East Tennessee State University East Tennessee State University
Digital Commons @ East Digital Commons @ East
Tennessee State University Tennessee State University
Electronic Theses and Dissertations Student Works
5-2019
A Better Predictor of NFL Success: Collegiate Performance or the A Better Predictor of NFL Success: Collegiate Performance or the
NFL Draft Combine? NFL Draft Combine?
Michael Gallagher
East Tennessee State University
Follow this and additional works at: https://dc.etsu.edu/etd
Part of the Sports Management Commons, and the Sports Studies Commons
Recommended Citation Recommended Citation
Gallagher, Michael, "A Better Predictor of NFL Success: Collegiate Performance or the NFL Draft
Combine?" (2019).
Electronic Theses and Dissertations.
Paper 3570. https://dc.etsu.edu/etd/3570
This Thesis - unrestricted is brought to you for free and open access by the Student Works at Digital Commons @
East Tennessee State University. It has been accepted for inclusion in Electronic Theses and Dissertations by an
authorized administrator of Digital Commons @ East Tennessee State University. For more information, please
contact [email protected].
A Better Predictor of NFL Success: Collegiate Performance or the NFL Draft Combine?
_____________________
A thesis
presented to
the faculty of the Department of Media and Communication
East Tennessee State University
In partial fulfillment
of the requirements for the degree
Master of Arts in Brand and Media Strategy
_____________________
by
Mike Gallagher
May 2019
_____________________
Dr. Susan E. Waters, Chair
Dr. Leslie McCallister
Dr. Adam Sayers
Keywords: NFL Draft, Combine, College Football, Power 5, Other 5
!
2!
ABSTRACT
A Better Predictor of NFL Success: Collegiate Performance or the NFL Draft Combine?
by
Mike Gallagher
NFL teams spend massive sums to ensure they are prepared for the future, but how
should they determine whom that future includes? This study set out to find what predicts
NFL success more accurately – collegiate in-game performance or the NFL Draft
Combine. In the sample of 2007-2012 first-round picks, 191 athletes were measured in
three NFL Draft Combine drills, two physical components, and a varying amount of in-
game collegiate and NFL performance statistical categories, dependent on position.
Secondarily, this work examined Power 5 and non-Power 5 players to determine if
attending a more prolific program was predictive of NFL success. Findings included that
40-yard dash and vertical jump are predictive of offensive linemen and cornerback NFL
success, that in-game collegiate statistics are most indicative of NFL success amongst
defensive players, and that Power 5 prospects are no more prepared for NFL success than
those coming from non-Power 5 schools.
!
3!
ACKNOWLEDGEMENTS
I never would have thought that getting a Masters degree was something that was
attainable for me, as no one in my family had even completed a Bachelors.
I owe everyone that supported me back in Minneapolis a big debt of gratitude for
keeping me positive throughout the 80- and 90-hour work/school weeks, specifically my
mother Kathy and my gal Meagan. Without your encouragement and reminders of my
goals, this would’ve been exorbitantly more difficult.
To Jay Sandos, the person responsible for hiring me for my graduate assistant
position in the athletics department, thanks for taking a chance on someone that had
never really accomplished much in broadcasting, and to Matt Higgins, the man that wrote
the recommendation that caught Jay’s eye, I hope you know how big of a deal that was in
my life. Thank you.
To my chair, Dr. Susan Waters, thank you for all the positive reinforcement.
Doing a thesis terrified me at first, but you have a way of making things seem more
manageable.
To Committee member Dr. Leslie McCallister, thank you for taking all the time
you did to help refine my paper and confirm my stats. Without the time you put in I
would’ve been a nervous wreck coming into defense day.
To my other Committee member Dr. Adam Sayers, thanks for taking the time out
of your busy schedule to get me to the finish line mate!
Finally, to my late father – thank you for instilling the work ethic in me I need to
see things through and give them my best, I wish you could be here today, and every day.
All of your efforts are greatly appreciated, thank you so much.
!
4!
TABLE OF CONTENTS
Page
ABSTRACT………….……………………………………………………………………2
ACKNOWLEDGEMENTS……………………………………………………………….3
LIST OF TABLES…..…………………………………………………………………….6
Chapter
1. INTRODUCTION…………………………………………..……………………..…..7
Building From The Outside In - Free Agency…………………………………….7
Building From Within – The NFL Draft……...………………………………….8
2. LITERATURE REVIEW…..………………….……………………………………..11
The NFL Draft – The Super Bowl for NFL Hopefuls………………...…………11
The History of the Combine…….…………….……………………………..…..13
The Combine – Effective or Irrational?……….….…….…………………..……13
Fails without Football Reason…….…………...………………………………...16
Combine Attempts from all Angles…..……....……………………………….…18
Standing Apart from Other Attempts….……...………………………………....23
Framing Theory…...……………………………………………………………..26
3. METHODS…..………………………………...……………………………………..33
Collegiate In-Game Statistics……………………….………………...…………35
NFL In-Game Statistics………………………………………………………….37
Combine Drills…………………………………………………………………...40
Mode of data Gathering……….…………………………………………............41
Reasoning for Collegiate & NFL Statistic Choices……………………………...42
!
5!
Sample Timeframe……………………………………………………………….45
“Power 5” Conferences v Non-“Power 5” Conferences…………………………46
4. RESULTS…………………..…………………….…………………………………..52
Positional Analysis.............……………………………………………………...54
Quarterbacks.............................................................................................54
Running Backs..........................................................................................57
Wide Receivers.........................................................................................60
Tight Ends.................................................................................................64
Offensive Linemen....................................................................................66
Defensive Tackles.....................................................................................71
Defensive Ends.........................................................................................75
Linebackers...............................................................................................79
Cornerbacks..............................................................................................82
Safeties......................................................................................................86
Power 5 & Big East, and Notre Dame v Non-Power 5.............................91
5. DISCUSSION…………...……………………...…………………………………98
Discussion of Findings.………………………………..………………………98
Research Question No. 1...……………………………..………………..99
Research Question No. 2…….…………………………………………108
Limitations, Direction for future Research……………………………………112
Conclusion…..……...…..…………………...…………………………………117
REFERENCES……..………………………………….……………..………………119
VITA………….………………………………………………………………………142
!
6!
LIST OF TABLES
Table Page
1. Collegiate & NFL Statistics Measured, by Position
.....................................
………
…
34
2. Collegiate In-Game Statistic Descriptions.................................………...…………….35
3. NFL In-Game Statistic Descriptions………………………………………...…….......38
4. NFL Draft Combine Drill Descriptions
…….….…………………...……
……
.....
40
5. Research Question No. 1 Significant P-Values – Offensive Positions…...........….....106
6. Research Question No. 2 Significant P-Values – Defensive Positions
.......................
107
7. Research Question No. 2 Significant P-Values – Power 5 v. Non-Power 5…..…....
..
111
!
!
!
!
!
!
!
!
!
!
!
7!
CHAPTER 1
INTRODUCTION
The National Football League is a dominant franchise across the United States,
drawing the lion’s share of U.S. sports fans’ consciousness and attention. In 2016, 33 of
the top 50 most-viewed sports broadcasts were NFL games (“2016 Ratings Wrap”, 2017),
headlined by the most-viewed program across all genres in 2016, Super Bowl 50, the 50
th
installment of the NFL’s annual championship game (Porter, 2016). The gap between the
Super Bowl’s audience, over 111 million total viewers, and the top non-NFL broadcast of
the year was over 70 million viewers, with the decisive Game 7 of the World Series
dwarfed by the annual spectacle (Porter, 2016).
With football drawing droves of individuals to their televisions and to events
throughout the year, large sums are paid from advertisers to the league, from the league
to the owners, and from the owners to the players. To entertain an audience the size the
NFL boasts takes healthy sums of money, and the importance of making sure that money
is given to the most talented players to maintain a successful product on the field takes
top talent evaluators and is of dire importance to a franchise’s success. When attempting
to find the right mix of individuals, those talent evaluators and general managers have
two options – the modern concept of free agency, which gives players the chance to move
around during their career, or the long-standing tradition of building through the NFL
Draft.
Building From The Outside In – Free Agency
First instituted in 1993 after more than 70 years of players being bound to their
original team in varying degrees (The history of, 2018), free agency marks the biggest
!
8!
structural change to the game in the last quarter century (Harrison, 2013), and has since
affected the trajectory of many different NFL franchises by making the danger of a player
leaving a franchise via unrestricted free agency all the more real. When this is the case,
the franchise receives no compensation, thus putting the player’s former franchise in a
difficult position. The San Diego Chargers were victims of this in 2006, letting free agent
quarterback Drew Brees test free agency waters, sign with New Orleans, and turn a 3-13
Saints team into a 10-6 group the next year, immediately leading them to an NFC
championship appearance and, a few seasons later, a Super Bowl win (Harrison, 2013).
While the Saints were winners in this one-sided transaction, as free agency often is, the
Chargers were left with Phillip Rivers, who has led the team to just one conference
championship appearance in 13 seasons.
Along with the empowerment of players that comes from the last quarter century
of free agency (Harrison, 2013), an ever-growing trend of players demanding a trade to a
different team he feels will fit him more properly, or hold out of football activities for a
new contract he believes is more fair, has arisen. Pittsburgh’s Le’Veon Bell did this
during the entire 2018 season, leaving the Steelers, a playoff team the previous four years
with Bell, on the outside looking in for the first time since 2013.
In the era of veteran players having more control of their futures, the more
traditional way of building a winning franchise, through the NFL Draft, has become that
much more vital (McCaffrey, 2015).
Building From Within – The NFL Draft
One of the most famous examples of the draft’s importance was a 2014 Green
Bay Packers team that had only two players on its 53-man roster that played for a
!
9!
different franchise to that point in their career. In that season the Packers were one of
eight teams left standing after winning a playoff game over the Washington Redskins.
General Manager Brian Gutekunst, in the Packers organization for 20 seasons, pointed
out that year was not an aberration, stating the Packers always have wanted to build
through the draft (Demovsky, 2018). The track record of success shows what is attainable
by doing so, as the Packers have finished at .500 or better each of the last 15 seasons,
finished with double-digit wins 12 times since Gutekunst joined the organization, while
winning Super Bow XLV in 2010, just the second 2
nd
Super Bowl in the last 50 years for
the team.
Aaron Rodgers, drafted by the Packers in the 2005 season, was the quarterback
that led Green Bay to the Super Bowl victory, though he was not the only draft product
that had Super Bowl success while playing for the same team that drafted him. In the last
16 NFL seasons, 11 of the Super Bowl Champions won with a quarterback that they
drafted, the only exceptions being Philadelphia’s Nick Foles, the aforementioned Brees in
New Orleans, a veteran Peyton Manning in Denver, and New York’s Eli Manning who
was all but a draftee of the Giants, traded on draft night from San Diego to New York in
2004.
Along with the aforementioned complications that can reveal themselves in free
agency and the realization that 31 other NFL teams may be competing for the same free
agents as a given organization, the NFL Draft presents a much more straightforward and
organic option for roster improvement. This makes the process of evaluating potential
draftees an important one, and the two main ways talent evaluators for NFL franchises do
so is by watching prospects during their college careers, and evaluating them at the NFL
!
10!
Draft Combine (Kelly, 2015). Abilities given the most preference differ from scout-to-
scout and franchise-to-franchise, which leaves much room for variance amongst opinion,
and a wandering question – is there one area in the evaluation process that acts as a
predictor of NFL success more than the others?
Previous research has looked at this issue in varying ways (Kuzmits & Adams,
2008; Robbins, 2010, Mulholland & Jensen, 2016; Park, 2016), though Lyons, Hoffman,
Michel, and Williams 2011 research “On the Predictive Efficiency of Past Performance
and Physical Ability: The Care of the National Football League” most mirrors the work
done in the current study.
While Lyons et al. (2011) provide a good baseline, the current study sets itself
apart from others by examining where the majority of money is spent in the draft, the
first-round selections, while also using a modern sample of a more representative size
that includes statistics over a longer period along with more in-depth analysis of
individual cases. In so doing, this work will examine first-round draftees from the 2007-
2012 years and their performance during their entire collegiate careers, in three drills at
the NFL Draft Combine, and throughout their seasons in the NFL. Following data
collection, statistical operations will be performed to determine if collegiate performance
or Combine performance are more predictive of success in the NFL. However, before this
study arrives at these conclusions, it is important to understand the maturation of the
Combine, stories of success and failure from it, as well as previous attempts to
statistically predict prospects prolific play from collegiate game performance.
!
11!
CHAPTER 2
LITERATURE REVIEW
In addition to being a tremendously important annual night for NFL teams, the
NFL Draft is also one of the most popular events outside of game action for football fans.
A three-night event, the draft gives NFL teams the chance to select collegiate prospects
that have declared themselves eligible for the selection process and demonstrated a
perceived ability to succeed as a professional. This event marries two dedicated and
sizeable fan bases, those of professional football and those of collegiate football, acting
as an anchor to keep enthusiasts’ attention from both worlds during a time where
football’s significance would otherwise be diminished. With the NFL being the giant of
the national sports scene and college football generating large numbers as well (over 25
million watched college football’s national title game in 2017), the 2017 NFL Draft saw
ESPN average nearly 7 million viewers during the draft’s first round, with another 1.3
million tuning in online. In addition, nearly 250,000 attended the event in the 2017 host
city of Philadelphia (Kay, 2017). The 4.0 rating that the draft garnered on its first night
was greater than the National Basketball Association’s playoff doubleheader the same
night (Kay, 2017). The NBA is widely thought of as No. 2 amongst the American sports
consumer, so the ratings battle that was won by the NFL that night should demonstrate
the size of the popularity gap between the NFL and other major sports.
The NFL Draft – The Super Bowl for NFL Hopefuls
With millions of eyes enamored with the three days that are set aside for the NFL
Draft, to think of the event as a standalone venture would be a mistake. Rather, the road
for potential draftees tends to loosely follow the one for professionals that have already
!
12!
made the NFL and are trying to reach the apex of their sport by getting to the Super
Bowl. For relative congruence, the NFL Draft is the Super Bowl for these individual
collegiate prospects – these are the nights their football journeys have been building
towards, much like the athletes that take the field for the NFL’s title game. Just like the
16 regular season games are lead ups to the NFL Playoffs which determine the
participants in the Super Bowl, the draft has events that lead into it as well, and much like
the NFL’s best must perform well to advance through the lead ups to the Super Bowl,
potential draftees must do the same as they approach the NFL Draft.
For instance, prospects play varying years of collegiate football, but must be at
least three years removed from high school in order to be drafted, so their game action is
plentiful prior to their draft eligibility. Think of this as the first step towards their Super
Bowl, much the same way the 16-game regular season is the first step for professionals to
reach the actual Super Bowl. The best teams after 16 games in the NFL are invited to the
playoffs, and should collegiate prospects stand out above their competition during their
time on the field in college, either for their physical gifts or exceptional statistical and
team production, or any combination of these factors, they can move on to the next step –
the “playoffs” of their road to the NFL. These playoffs are also by invite only and give
prospects the chance to enter the collective consciousness of the professional football
world. The similarities continue – the stakes are high, there are winners and losers, there
are sweat and tears, and for those that are successful, millions of dollars and a chance at
incredible fame. For those that fail, dreams can be dashed.
All that can be said of the Super Bowl, and so too of the final step before
prospects hear their name called at the Draft, or live through the painful opposite. The
!
13!
place some believe future champions are born and franchise cornerstones are made, the
playoffs of these prospects’ livelihood - the NFL Scouting Combine.
The History of The Combine
The Combine comes from humble beginnings, starting in 1982 with 163 athletes
participating. At that time, the event was called the National Invitational Camp put on by
National Football Scouting, and was largely held as a way to ascertain medical
information for the prospects that attended (“History”, 2017). In 1985, National Football
Scouting merged its camp with two others, creating the first scouting camp in which all
28 teams that were part of the NFL at the time attended. The event was renamed the NFL
Scouting Combine post-merger, and began measuring prospective NFL athletes in drills
such as the 40-yard dash, the bench press, and the vertical leap. In 1987, the Combine
was moved to its current home of Indianapolis (Gabriel, 2017), and has grown into a
media event that is widely covered and enthusiastically consumed. The first Combine to
be televised was in 2004 with the inception of the NFL Network (Wood, 2004), and the
event has grown to over 300 athletes on a yearly basis. The event contains 14 different
mental and physical drills, and for the first time in 2017 included spectators (“NFL
announces”, 2016). This holy grail of scouting venues is watched with wonder by fans,
flocked to by scouts, and highly sought after by collegiate prospects. Opinions are formed
and conclusions are drawn based solely off the performances put forth on this big stage.
The Combine – Effective or Irrational?
The effectiveness of the Combine, though, has been a point of contention since its
rise to prominence, and while it may possess some measure of face validity for future
NFL performance (Kuzmits & Adams, 2008), just how much of a measuring stick it can
!
14!
be is a point that is certainly in doubt. The 40-yard dash measures in-line, unimpeded
speed for 40 yards, the bench press logs how many times a player can push 225 pounds in
the air while lying on his back, and the vertical leap records a player standing still and
jumping in the air as high as he can. These are not activities often seen on an NFL field,
which has led to many misnomers regarding future success of prospects.
One of the most famous examples of misidentifying a player’s ability at the NFL
draft based off one of the battery of tests at the Combine was draft prospect Terrell
Suggs. In Suggs’ final season at Arizona State in 2002 he recorded an NCAA record 24
sacks in a season, and still ranks as the NCAA Division I all-time leader in career sacks
with 44 (“Football Bowl”, 2017). Suggs ran a 4.83 40-yard dash at the Combine and
repped 225 pounds only 19 times on the bench press, both numbers considered subpar for
a top prospect at his position. Suggs saw his draft stock fall thanks to these valuations
(Silverman, 2012), but would go on to be one of the most dominant defensive players in
the NFL in his playing career, winning Defensive Rookie of the Year in his first pro
season, Defensive Player of the Year in 2011, and being selected to six Pro Bowls during
a 15-year NFL career that is still active as of the conclusion of this study.
Various Combine measurables for Joe Haden, Anquan Boldin, and Tom Brady in
which they did not score well all led to those prospects tumbling down draft boards in the
eyes of some, but the teams that drafted them trusted game footage more than the
numbers from the Combine, and all have made at least one Pro Bowl, making good on
the team’s trust in their ability on the football field (Ruiz, 2016). For example, the
aforementioned Brady’s 40-yard dash time was 5.28 when he was drafted in 2000, the
slowest amongst active NFL quarterbacks as of the 2016 season (Duffy, 2017). Brady
!
15!
went on to win six Super Bowls and be named to the Pro Bowl 12 times, and is still one
of the league’s best quarterbacks at 41 years of age. These types of stories led Lyons et
al. (2011) to suggest a new testing battery that is less about judging indicators of success
and more about setting up scenarios that may actually occur within a football game.
Many players have performed well at the Combine, despite college careers that
lacked consistency and reliability. Matt Jones, a quarterback at Arkansas in college, was
not considered one of the top quarterbacks coming out of the collegiate ranks that year, so
he was drafted as a wide receiver by the Jacksonville Jaguars in the first round in 2005 –
a result of his performance in the 40-yard dash and vertical jump, considered by most to
be indicative of wide receiver success. Jones started just 15 games for Jacksonville, not
appearing on an NFL roster again after the 2008 season (“Matt Jones”, 2009). Numbers
from the Combine predicted great success positionally for Jones, but having never
appeared in a game at receiver for Arkansas, the Jaguars had little tangible backing for
this selection.
Troy Williamson was selected seventh in the same draft as Jones by the
Minnesota Vikings following his best season for South Carolina, which netted
Williamson just 835 receiving yards, a mark that put him outside the top 40 in the
country that season (“FBS Player”, 2006). After running a 4.34 40-yard dash at the
Combine in 2005, fifth-fastest amongst over 300 athletes in attendance, he was taken by
the Vikings, only to be out of the NFL after playing in just 49 games over 5 seasons
(“Troy Williamson”, 2010).
Mike Mamula, drafted seventh overall in 1995 by the Philadelphia Eagles, was
projected as a third-round pick off of his game performance entering the Combine. At
!
16!
6’4”, 252 pounds, Mamula ran a 4.58 40-yard dash, leapt nearly 40 inches, and produced
a combine-high 37 reps on the bench press (Breer, 2015), leading the Eagles to select
him, as it proved to be, much too early, as he lasted only five years in the NFL.
The latter three examples of Jones, Williamson, and Mamula being selected off of
Combine measurements rather than game performance lasted a combined 14 years in the
NFL, 4.66 per athlete, half the average career of an NFL first-round pick (“Average
playing career”, 2017).
Fails without Football Reason
While these examples seem telling, they cover only a few of the stories from the
Combine, and truly, one can find examples on all sides of a subject that may be outliers.
Some players that have gone through their collegiate journey and reached the pinnacle of
being drafted in the first round have encountered outside obstacles that have thrown them
off course that have nothing to do with their collegiate, or Combine, performances.
The Oakland Raiders made what is widely considered one of the worst selections
in the history of the NFL Draft, taking Jamarcus Russell first overall in 2007, only to see
him play in just 31 NFL games. Russell’s impressive stature, standing 6’5” and weighing
in at 265 pounds, along with his ability to wow throwing the football long distances in
non-game situations (Evolve IMG, 2013), overrode an incomplete collegiate resume that
lacked ample performance to justify a selection at No. 1 overall, and also a work ethic
that was not at the level it needed to be, partially because of an alleged sleep apnea issue
(Wertheim, 2011).
Charles Rodgers, a wide receiver out of Michigan State selected No. 2 overall in
the 2003 NFL Draft who some peers called the best athlete they had ever seen (Tucker,
!
17!
2017), had an impressive collegiate career, putting forth back-to-back 1,300-yard seasons
in college to go along with 27 combined touchdowns those two years (“Charles Rogers”,
2003). Rodgers also ran one of the best 40-yard dash times in NFL Combine history,
logging a 4.28 (Fransen, 2017) while measuring 6’3” and 220 pounds, a rare combination
of size and speed at the position. But Rogers broke his clavicle in back-to-back seasons,
became addicted to pain pills, and as of April 2017, still admits to smoking marijuana
every day (Tucker, 2017), a drug outlawed in the NFL’s Substance Abuse Policy which
Rodgers was suspended for violating in 2005. Rogers later found legal trouble for his
marijuana usage as well (“Judge Issues Warrant”, 2013).
David Carr, who was selected by the Houston Texans first overall in the 2002
NFL Draft, built and developed his way to a huge senior season, throwing for over 4800
yards and 46 touchdowns to just nine interceptions at Fresno State (“David Carr”, 2002).
He ran a 4.67 40-yard dash, a virtual tie for fastest amongst quarterbacks at the Combine
in 2002, leapt 35 inches, and stood nearly 6’4”, 225 pounds (“2002 NFL Combine”,
2002), ideal quarterback size. But 2002 was the Texans inaugural year in the NFL,
entering the league via expansion, and their first pick in the expansion draft, Tony
Boselli, an offensive tackle taken to protect Carr, never played a game because of
injuries. Partially due to Boselli’s absence, Carr would be sacked 249 times in his first
five seasons in the NFL, and still holds two of the top three marks for most times sacked
in a season in NFL history (“NFL Sacked”, 2017). Carr’s record as a starter was 23-52
with Houston, and he would be released after five years with the team.
!
18!
Combine Attempts from all Angles
Many tales of failure and success out of the Combine can be told, and calling out
individual stories can be effective to a point, but what this study looks for is consistency
in order to determine the best method to judge a collegiate prospect’s professional
trajectory, something that to this date has been fleeting and fruitless for those that have
attempted to find it.
One example of an effort to standardize a statistical system of evaluation came
from Lopez (2010), who isolated the quarterback position with the belief that the 26-27-
60 formula could help determine a future NFL signal-caller’s worth. This formula states
that if a quarterback scores 26 on the general aptitude test at the Combine (called the
Wonderlic), starts at least 27 games in his collegiate career, and completes at least 60
percent of passes in college, he will be successful in the NFL. Some examples of this
formula holding true include NFL mainstays Peyton Manning, Phillip Rivers, Eli
Manning, Drew Brees, Tony Romo, Matt Ryan, and Matt Stafford, at least two of which
(P. Manning and Drew Brees) will one day be Hall of Famers, while four more (E.
Manning, Ryan, Stafford, Romo) have a legitimate chance to join them. Those that have
not met the formula include the aforementioned David Carr and Jamarcus Russell, Tim
Couch, Akili Smith, Joey Harrington, Michael Vick, Ryan Leaf, Daunte Culpepper, and
Vince Young. Five of those quarterbacks are ranked amongst the 15 worst of all-time
(Beckett, 2015), while another spent time in jail (“Michael Vick”, 2015) and another
faced alleged mental health barriers (Travis, 2008), none having experienced consistent
success. Still, this simplistic rule has its exceptions, as the same Tom Brady referenced
earlier did not fit the criteria but has amassed one of the greatest careers in NFL history,
!
19!
while Super Bowl winners and long-time starters for their teams Ben Roethlisberger and
Joe Flacco also came up short of meeting the 26-27-60 threshold but thrived regardless.
While Lopez was not successful in his collegiate and Combine combination
approach, previous studies have had success in showing that 40-yard dash times produced
at the Combine can be representative of future NFL success, but the notion only holds on
a position-specific basis.
Kuzmits and Adams (2008) supported the above, studying wide receivers, running
backs, and quarterbacks from 1999-2004 across 10 drills at the Combine, putting together
correlations based off 10 variables measured against those drills. Of the 218 correlations
they calculated, only 14 were significant, and in the case of the quarterback and wide
receiver groups, the significance was indicative of a random chance model, therefore the
authors dismissed the findings as irrelevant. The only significant correlation that held
after the authors scrutinized the data further was the relationship between running back
sprint times and NFL success. All other variables across positions did not appear to hold
up under the statistical spotlight, shedding skepticism on an overwhelming majority of
Combine predictability.
Mulholland and Jensen (2016) did an extensive study of wide receivers’ college
vs. Combine data using regression analyses and recursive partitioning decision trees.
Across the 15 years included in their study, the one predictor of draft success from the
Combine that was deemed significant was the 40-yard dash. They also collaborated on a
similar work in which tight ends were the subject, the outcome showing players were
selected in the draft based on 40-yard dash and vertical jump.
!
20!
Park (2016) had different means to come to the same conclusion, using multi-
linear regression and random forest statistical regression to find that 85% of wide
receivers that were one standard deviation above the mean in the 40-yard dash were
drafted, while only 30% that were one standard deviation below were taken.
However, an important distinction must be noted. While 40-yard dash was
indicative of being drafted (Park, 2016; Mulholland & Jensen, 2016), it did not predict
NFL success statistically in Park’s study and was not in the most prevalent predictor
group in Mulholland & Jensen’s (2016) study. The current study does not look to
discover what numbers lead to a player getting taken in the draft, but what numbers lead
to a player having success once drafted.
Robbins (2010) attempted to draw a connection between normalizing data to
better design a model that would be reliable for predicting players selection in the NFL
draft. While that aspect of the study failed, as no connection was found between
normalizing vs. using raw or ratio-scaled Combine data in order to predict draft order, it
found that 40-yard dash and vertical leap are the two most significant drills that lead to
NFL teams drafting players across positions.
The information gleaned across these studies is extremely valuable, as it shows
that 40-yard dash time most often influences scouts and decision-makers to draft a player,
the point driven home by the Robbins article. In total, of the articles listed above, four set
out to find predictive success of players in the NFL based off their Combine
measurements (Kuzmits & Adams, 2008; Park, 2016; Mulholland & Jensen, 2014 &
2016), while Robbins (2010) attempted to normalize Combine data for predictive
purposes. Of the five studies, Kuzmits and Adams (2008), Park (2016), and both
!
21!
Mulholland & Jensen (2014, 2016) studies are most applicable to this work. Kuzmits and
Adams (2008) supported the 40-yard dash in predicting NFL success for running backs,
Park (2016) supported it in predicting draft positioning for running backs and wide
receivers, and Mullholland & Jensen (2016) mildly supported it for predicting NFL
success amongst NFL tight ends and wide receivers. All four, though, reject the Combine
in predicting NFL success in every other area. These mixed results, along with the
statement from the Robbins (2010) study that found NFL brass use 40-yard dash and
vertical leap amongst Combine drills most-often in selecting prospects, has led the
authors to look at the 40-yard dash and vertical jump as two drills to examine potential
relationships between NFL success and quality performance in the drills. Because of
results like Mike Mamula and Terrell Suggs described earlier, bench press is the third
drill studied in this work to attempt to determine whether the Combine is an accurate
predictor of success across positions.
Most related to this work is the journey Lyons et al. (2011) embarked on, pitting
what they called “signs” (generalized tests of ability, in both of our cases the Combine)
against “samples” (directly related prior experiences, in both of our cases collegiate game
performance) in hopes of finding which of the two resulted in higher criterion-related
validity. Lyons et al.’s (2011) study differs from the previous examples discussed directly
above as it did not exclude the majority of positions (Kuzmits & Adams, 2008; Park,
2016; Mulholland & Jensen, 2014 & 2016), did not normalize and scale data (Robbins,
2010), and did not attempt to determine draft success (Park, 2016; Mulholland & Jensen,
2014& 2016; Robbins, 2010), only future NFL success, all of which can also be said
about the current study as well. The results of Lyons et al. (2011) study were supportive
!
22!
of previously presented data in terms of “signs”, as the 40-yard dash was the most
statistically significant relationship of Combine drills and best predictor of NFL success
after thorough intercorelational and hierarchical regression analyses. But while the 40-
yard dash was the best predictor amongst “signs”, it, as well as the other drills that were
examined at the Combine, were dwarfed by the results when judging collegiate
performance, which was deemed by a wide margin the best predictor of NFL success of
the variables measured. While Lyons et al. (2011) bolstered results mentioned in the four
studies above, their study also presents guidelines to answer of one of our most difficult
questions in this study – how does one measure successful collegiate performance?
Lyons et al. (2011) did this a few ways to ensure they were capturing all of the
outliers that could be involved in measuring collegiate performance. They started by
including a control variable that specified the player’s affiliation, or lack thereof, with a
Bowl Championship Series conference during college. This was done because, even
within Division I competition, which is broken up into clusters called conferences,
different perceived skill levels are present depending on which conference an athlete is a
part of. Most important amongst these conferences, according to Lyons et al. (2011), is an
affiliation with a conference within the BCS, historically the highest level of college
football competition prior to its elimination as an entity.
Following the institution of a control variable, the authors standardized their
college performance data to calculate averages from each draft they examined (three from
2002-2004), and determined how those averages stood up to criterion-related validity by
testing different standard deviations removed from the average. Those averages came
from 11 different statistical categories for offensive players and five for defensive
!
23!
players, of which the categories were broken down positionally after they were separated
into offensive and defensive subsets. The statistical categories that were tallied for
collegiate performance were also used by the authors to gather NFL statistics from the
first four years of the players’ professional career, ensuring consistency across data
collected, and also acting as a way to log college vs. professional performance in a
directly representative way.
Standing Apart from Other Attempts
While Lyons et al. (2011) prepared a thorough study that gave many queues for
how to carry out this work, certain limitations to their methods exist, and the present
study seeks to address these limitations and refine their model. Specifically, their control
variable lacked necessary depth to identify trends amongst conferences, their work
ignored the wide variance of skill and importance between the subgroups of first-round
draftees and the rest of those selected, they took only partial data from subjects collegiate
and professional careers, their sample covered only three years, and their data is now
outdated, coming from a time range in which the Combine was just becoming the event
that it is now and scouting practices were just beginning to evolve.
Firstly, while instituting a control variable was a wise idea to ensure differences in
competition throughout BCS schools and the rest of college football, the method leaves
questions unanswered. Most notably, while players of BCS schools may have more, or
less success based on the competition they face and their natural ability that has landed
them a scholarship at a large, successful program, there are many subtleties across
different conferences, and further, schools. These differences and subtleties may skew
data one way and make BCS or non-BCS schools look attractive, when in fact they
!
24!
simply look that way because a few schools from the entire BCS, or a few from each
conference, are weighting data and overlooking a lack of consistency. There is room for
misinterpretation by coding for only two options, BCS and non-BCS, so the current study
codes for each school to attempt to determine if attending a certain school in a
conference, or the conference as a whole, gives prospects a better chance at professional
prosperity.
Next, they did not separate draft prospects into different subgroups, despite the
fact that the NFL does so with the round that a player is selected in. Rather than taking
queues from the organization, Lyons et al. (2011) studied all seven rounds of the NFL
Draft over their three sampled seasons. Because this study hopes to act as a guide to NFL
teams on where to effectively delegate the most lucrative contracts they sign every
spring, the author’s focus is only on first-round picks. Gaines and Yuhari (2017) showed
via graphs the steep drop off and relatively small amount of money spent past round one
of the Draft. The most telling image was a graph closely resembling the second half of a
bell curve with first-round pick compensation at the top left of the one quadrant graph, a
quick decline moving from left to right with seventh-round picks on the bottom right of
the X axis. This illustrates why, on top of the fact that it is important to compare like-
samples for accurate information, this study logs only first round draftee activities.
Further, their 2011 study looked only at the final year of a player’s statistics in
college, but to validate that a player had consistent success collegiately this study
includes data from a prospect’s entire collegiate career. This should help ensure that the
author has a large enough sample size of the players’ collegiate performances to
disregard any one-off seasons that could skew results.
!
25!
Similarly, the foundational study for this work only looked at the first four years
of NFL performance amongst subjects. This study will look at the entire career of
participants to judge longevity and overall impact throughout their professional presence.
Some in this study, particularly those that turned professional in 2011 and 2012, will not
have complete career data to count, as many of the first-round picks are still active in the
NFL. Despite this fact, a more complete subject snapshot can be developed based on the
larger amount of data this work will collect from a subject’s career, and those still playing
can, at the most elementary level, already be considered successful NFL players, as the
majority that are still active have already eclipsed the sample’s average for NFL first-
round pick career length at their respective positions.
Finally, the 2002-2004 time-period represents multiple issues. First, it is only
three years, half the length of the current study. Second, being a full 15 years ago, the
current work is of a more modern nature, with 2002-2004 no longer representing the
current landscape of scouting practices by organizations and preparation by players for
the Combine. With rapid advancements in technology, strength programs, desired
attributes, and an ever-evolving definition of the modern NFL athlete, let alone the
evolution of the event itself (Fierro, 2015), the player data may be rendered useless in
comparison to near-current day measurements in the modern age of scouting. Add that to
the growing sample of successful, and unsuccessful, draft picks made by each franchise
that grows the sample and weighs on the psyche of decision-makers selecting or not
selecting certain players with their next pick, and definitions of how to draft look
completely different than they did 15 years ago.
!
26!
Framing Theory
Originally founded in 1982, the NFL Draft Combine was meant to be a more
convenient means of getting medical information on prospective pros (National Scouting
Combine, 2017). In the late-1970s and early-1980s, it became common practice for
players to fly to teams and have one-on-one meetings. It worked well for the teams, but
the draft hopefuls would have to do all the traveling, leaving many exhausted by the
conclusion of the process. One story of this being the case took on a particular
significance towards progress and change being initiated, a tale told by scouting legend
Gil Brandt:
“It all came about because of Nolan Cromwell. Nolan Cromwell was a Kansas
running back, that was a great player and got hurt,” Brandt recalled. “He was
flying around the country and we were like the 14th stop, we the Cowboys, and he
had all these envelopes under his arm, the guy looked all tired out. [Cowboys
president and general manager] Tex Schramm and I happened to meet him on the
elevator coming up to our office. ‘Where’ve y’all been?’ ‘Oh, I’ve been to Seattle,
I’ve been here, I’ve been there.’ Tex said, ‘There’s got to be a better way.’”
(Knaak, 2018, para. 10)
Schramm was one of the pioneers of bringing the Combine into existence, with
this story playing a factor in the steps he would later take in modernizing and
streamlining the idea. Schramm’s Cowboys held a mini-Combine in the early-1980s with
three other teams, Brandt confirming that the idea was to have prospects have to go
through medical examinations only one time, and the idea caught fire, expanding to the
entire league just a couple of days later (Knaak, 2018).
!
27!
So how did medical exams and interviews turn into 40-yard dashes and bench
press reps? Framing.
The psychological theory was developed by Gregory Bateson in 1972, stating
framing was a “spatial and temporary bounding of set of interactive messages” (Arowolo,
2017). Two years later, Erving Goffman repurposed Bateson’s Framing Theory for
sociology, introducing Primary, Natural, and Social Frameworks, offering that people
interpret what is going on around their world through their frameworks (Goffman, 1974).
Further classifying Framing Theory, McCombs et al. (1997) found direct correlation
between the news media’s agenda-setting with political candidates, establishing framing
as “second-level agenda setting”, later defined by Balmas and Sheafer (2010) as the
media telling us how to think about something, a step further than first-level agenda
setting, which tells us what to think about.
While framing is not studied in sports as often as it is in politics, it exists on a
game-in, game-out basis. Play-by-play broadcasters open their pregame monologues with
first-level agenda setting, introducing the big storylines of the night to come. Throughout
the game the analyst, sideline reporter, and others that may fill out the announcing crew
give in-depth analysis at the second level of agenda setting, telling the viewer how to
think about what they are seeing or listening to. In the print and digital media realm,
columnists and beat writers deliver the same type of coverage on social media, websites,
and in the various newspaper in which they are published.
While games draw the attention listed above, one doesn’t have to look far to find
the intersection of both levels of agenda-setting and the NFL Combine, as its popularity
spike over the last 15 years came at roughly the same time media attention reached a new
!
28!
milestone in 2004. That year, with the inception of the NFL Network, a television
channel dedicated completely to National Football League programming, the decision
was made to broadcast the Combine on linear cable TV for the first time in the event’s
history (Wood, 2004). Since the initial broadcast, the Combine has adjusted its schedule
to better cater to the consumer’s needs, moving to Friday-Monday in 2015 from a
Saturday-Tuesday schedule in prior years, a positioning tact that logged the top single-
day viewership in event history (“Saturday’s NFL Network”, 2015). In 2017, the event
had its best overall showing, while 2018’s viewership numbers were 19 percent higher
than 2016’s (Paulsen, 2018).
The NFL Network and NFL Media, the two parties responsible for the coverage
of the event on television, were unlikely to have the kind of success they have by
focusing on medical examinations and team interviews, so they needed to do some first-
level agenda setting, as well as second-level agenda setting in order to put a successful
media product on their airwaves. Considering NFL Network shows very few live football
games because of its rights agreements with FOX, ESPN, NBC, CBS, etc., this task was
of incredible import as NFL Network executives often described the Combine as one of
their network’s “tent poles” (Ourand, 2017), media-speak for the broadcasts being
foundational in the success of their channel.
The first-level agenda setting involved identifying drills that had previously
garnered interest and attention from the public, as well as played a factor in determining
success or “draftability” of prospects. With the instant-gratification nature of today’s
society, attention spans shrinking to under eight seconds in the average American
(“Microsoft attention spans”, 2015), the fit could not have been better for the 40-yard
!
29!
dash, which lasts from 4.2 to 5.5 seconds depending on position of the player running it.
The vertical leap and bench press, the other two events focused on in this study, also are
quick in nature, with all three drills mentioned playing a significant role in the
“draftability” of the players, per research put forth in the literature review portion of this
work. Other televised drills include the 3-Cone, Broad Jump, and a score of others that
keep viewers attention by not demanding it for very long.
With the NFL Network and NFL Media showing, and telling, us what to pay
attention to, their next task was to make the coverage relevant to interested viewers by
telling them how to think about the performances by framing the topic in the lens they
desired. This applies to the over 1,000 media members gathered in Indianapolis for the
Combine as well, who take different stories that come out of the Combine and deliver
them to their audiences under the guise of different frames.
John Ross is a recent example of framing with his 40-yard dash performance. In
2017, Ross broke the modern Combine record in the drill, running a 4.22, beating the
previous record of 4.24 held by Rondel Menendez and Chris Johnson. NFL Network host
Rich Eisen watched in amazement: “4.22. No Way! According to what we just timed,
Chris Johnson is no longer the record holder.” Another analyst chimed in: “I would go to
the house, he can borrow my jet and go home.” (White, 2017, sec. 21).
By immortalizing Menendez and Johnson’s 4.24 40-yard dash the way the media
did, breaking the record brought Ross from another face in the crowd of wide receivers
hoping to be first-round selections into a household name with major expectations and
excitement placed around him, despite the fact that previous record holders have spotty
track records of NFL success. Of those that have run 4.30 or below in the last 20 years,
!
30!
only two would be considered booming successes, those being long-time Pro Bowl
cornerback Champ Bailey, the other being one of just six players to rush for over 2,000
yards in a single season, Chris Johnson. The 17 others (Fastest 40, 2017) with a time of
4.30 or below include Menendez, Dri Archer, and Josh Robinson, who were in and out of
the league in just a few seasons, while many others have managed to continue their
professional careers despite repeated unproductive years on the field. Early returns on
Ross support the latter type of result rather than the Johnson or Bailey career path, as he
has only caught 21 passes in the NFL since being drafted ninth overall in 2017 by
Cincinnati.
Other examples of framing by the media effecting public perception include
Byron Jones, who was rated the No. 25 cornerback in the 2015 Draft (Manfred, 2015)
prior to the Combine. At the event, however, Jones leapt 44.5 inches in the vertical jump,
NFL Analyst Mike Mayock stating that Jones was “gifted genetically” and that people
need to pay more attention to him (NFL, 2017). Though Mayock’s comments dismissed
any hard work that Jones may have put into achieving such a top-end result, fans and
talent evaluators did take notice after Jones performance garnered national attention, and
he would end up as the fourth cornerback taken in the Draft. So far his career has featured
a position change and three seasons that have fans of the Dallas Cowboys, the team he
was selected by, waiting for Jones to make a major impact (Kohut, 2018).
Another recent example is current Jacksonville Jaguars cornerback Jalen Ramsey.
In reviewing his work throughout the 2016 Combine drills, Mayock said Ramsey is
“super athletic” and an “alpha male”. Eisen then added that he was “sold” on Ramsey as
a prospect after a drill that showed Ramsey on a field by himself catching a football and
!
31!
returning it for a touchdown. Eisen then haphazardly compared Ramsey to Patrick
Peterson during the drill, who has made seven consecutive Pro Bowls in the NFL,
Ramsey at the time having played a total of zero snaps in pro football (NFL, 2015). These
frames are meant to help the viewer place a player next to someone they are more
familiar with, classifying Ramsey with someone that has been through the Combine
process before and has showcased similar ability, but it paints an unrealistic and
deceptive picture in a situation that showcases little about in-game performance. Making
non-sequitir assessments of one’s ability can mislead the public, and scouts as well,
something that the Combine and its media have arguably been doing for years (Caputo,
2010).
The media framing of performances at the Combine largely represents and
codifies the overall criticism of the event, with commentators focusing on athleticism,
“genetic gifts”, and largely meaningless numbers. Much like the drills themselves, the
media focus very little on how these drills and the athleticism shown by participants can
translate to football success and what one has to do with another, ignoring that many of
the drills will never come into play on a football field during a game, and that many of
the top performers in the drills have gone on to substandard NFL careers.
With the above in mind, this work sets out to determine the answer to these two
research questions:
RQ1: Which is the better predictor of NFL success – prospects in-game collegiate
performance, or the three most scrutinized Combine drills - 40-yard dash, bench
press, and vertical leap?
!
32!
RQ2: Is there a significant advantage in drafting prospects from “Power 5”
collegiate programs over “Other 5”, FCS, Division II, or Division III programs?
!
33!
CHAPTER 3
METHODS
This study will examine three data sets, beginning with the NFL Combine. Data
from the NFL Combine will include the 40-yard dash, bench press, and vertical leap of
NFL Combine participants that were selected as first-round picks from 2007-2012. This
data will be analyzed to determine if correlations exist between the drills performed and
NFL success. This will be conducted positionally on both the offensive and defensive
sides of the ball. According to previous research, the 40-yard dash is the most
representative drill of NFL success (Kuzmits & Adams, 2008; Mulholland & Jensen,
2016; Park, 2016); 40-yard dash and vertical leap are the drills most often considered by
NFL decision makers when drafting players (Mulholland & Jensen, 2016; Park, 2016;
Robbins, 2010); and research performed in this study shows bench press is consistent
with over- or under-valuing prospects.
The second data set will be collegiate statistical performance. Lyons et al. (2011)
did not divulge which statistical categories they measured for collegiate, or NFL,
performance, but this work will draw on the criterion different college, and NFL, position
players are most often measured on via the five most popular sports websites according to
Alexa’s global traffic ranking. The websites are YahooSports.com, ESPN.com
BleacherReport.com, CBSSports.com, and SportsIllustrated.com (“Top 15 most”, 2018).
Additionally, prior to data collection, it was discovered that over 95 percent of the
websites used in data collection and fact-checking logged all of these defensive stats,
those totals coming from software data-entry programs that have partnerships with
institutions and large governing organizations like the NCAA and NFL. Because of the
!
34!
deep-rooted nature of these statistics in the data-technology sector, and the obvious
recognition by the NCAA and NFL of the importance of these statistics simply indicated
by their partnerships with these statistical groups, there is no worry that data is being
missed in collection that could be valuable in predicting success.
The third and final data set will be NFL statistics of the collegiate prospects, their
entire career statistics collected and then cross-examined against Combine and collegiate
numbers. The sample will be broken into 10 positional data sets with varying position-
specific measurements determining in-game success, while all 191 athletes in the sample
will also be judged off of height, weight, and years played. The collegiate and NFL data
categories are as follows:
Table 1.
Collegiate & NFL Statistics Measured, by Position
POSITION STATISTICAL MEASURE
Quarterback Games Played, Completion, Percentage,
Passing Yards, Yards Per Attempt,
Rushing Yards, Touchdowns
Running Back Games Played, Rushing Yards,
Receiving Yards, Yards Per Carry,
Touchdowns
Wide Receiver Games Played, Receptions, Receiving
Yards, Yards Per Catch,
Touchdowns
Tight End Games Played, Receptions, Receiving
Yards, Yards Per Catch,
Touchdowns
Offensive Lineman Games Played, Pro Bowls, All-Pro
Seasons
Defensive Tackle Games Played, Tackles, Sacks, Tackles
!
35!
For Loss, Fumbles Forced, Fumbles
Recovered, Interceptions, Passes
Defensed
Defensive End Games Played, Tackles, Sacks, Tackles
For Loss, Fumbles Forced, Fumbles
Recovered, Interceptions, Passes
Defensed
Linebacker Games Played, Tackles, Sacks, Tackles
For Loss, Fumbles Forced, Fumbles
Recovered, Interceptions, Passes
Defensed
Cornerback Games Played, Tackles, Sacks, Tackles
For Loss, Fumbles Forced, Fumbles
Recovered, Interceptions, Passes
Defensed
Safety Games Played, Tackles, Sacks, Tackles
For Loss, Fumbles Forced, Fumbles
Recovered, Interceptions, Passes
Defensed
Collegiate In-Game Statistics
The statistics for collegiate game performance will be collected from sports-
reference.com, the most comprehensive sports statistics database on the Internet and the
2013 Sloan Sports Conference award winner for Best Analytics Innovation/Technology.
When additional resources are needed to verify statistics that may be unavailable from
sports-reference.com, college football program’s team websites will be consulted.
Table 2.
Collegiate In-Game Statistic Descriptions
STATISTIC DESCRIPTION
Years Played The number of years the subject played
at his final Division I or FCS
institution.
Games Played The number of games played by the
36!
subject throughout his Division I
college career.
Completion Percentage For quarterbacks, the percentage of
passes completed throughout the
subject’s FCS or Division I career.
Passing Yards For quarterbacks, the number of passing
yards accrued throughout the
subject’s FCS or Division I career.
Yards Per Attempt For quarterbacks, the number of yards
gained per passing attempt
throughout the subject’s FCS or
Division I career.
Rushing Yards For quarterbacks and running backs, the
number of yards gained running the
football during the subject’s FCS or
Division I career.
Touchdowns For quarterbacks, running backs, wide
receivers, and tight ends, the
touchdowns from scrimmage
(passing, rushing, receiving) scored
throughout the subject’s FCS or
Division I career.
Yards Per Carry For running backs, the yards averaged
per running play throughout the
subject’s FCS or Division I career.
Receiving Yards For running backs, wide receivers, and
tight ends, the number or yards
gained as the recipient of a pass from
another player throughout the
subject’s FCS or Division I career.
Yards Per Catch For wide receivers and tight ends, the
average number of yards gained per
reception throughout the subject’s
FCS or Division I career.
Tackles For all defensive players, the amount of
total tackles made throughout the
subject’s FCS or Division I career.
37!
Tackles For Loss For all defensive players, the amount of
tackles behind the line of scrimmage
made throughout the subject’s FCS
or Division I career.
Sacks For all defensive players, the amount of
times the subject tackled the
quarterback behind the line of
scrimmage on a pass play in the
subject’s FCS or Division I career.
Forced Fumbles For all defensive players, the amount of
times the subject stripped the ball
from an opposing ball carrier
throughout the subject’s FCS or
Division I career.
Fumble Recoveries For all defensive players, the amount of
forced fumbles possessed to
complete a turnover throughout the
subject’s FCS or Division I career.
Interceptions For all defensive players, the amount of
passes caught to force a turnover
throughout the subject’s FCS or
Division I career.
Passes Defensed For all defensive players, the amount of
pass breakups plus the number of
passes intercepted throughout the
subject’s FCS or Division I career.
NFL In-Game Statistics
When collecting NFL statistics, Pro-Football-Reference.com will be the source
for collection of data, under the umbrella of the aforementioned Sports-Reference.com,
the company named in 2010 as having one of the top 50 websites in the world by TIME
Magazine (Staff, 2010). On the rare occasion that a statistic may not available,
ESPN.com will be consulted.
38!
Table 3.
NFL In-Game Statistic Descriptions
STATISTIC DESCRIPTION
Years Played The number of years the subject played
in the NFL.
Games Played The number of games played by the
subject throughout their NFL career.
Completion Percentage For quarterbacks, the percentage of
passes completed throughout the
subject’s NFL career.
Passing Yards For quarterbacks, the number of passing
yards accrued throughout the
subject’s NFL career.
Yards Per Attempt For quarterbacks, the number of yards
gained per passing attempt
throughout the subject’s NFL career.
Rushing Yards For quarterbacks and running backs, the
number of yards gained running the
football during the subject’s NFL
career.
Touchdowns For quarterbacks, running backs, wide
receivers, and tight ends, the
touchdowns from scrimmage
(passing, rushing, receiving) scored
throughout the subject’s NFL career.
Yards Per Carry For running backs, the yards averaged
per running play throughout the
subject’s NFL career.
Receiving Yards For running backs, wide receivers, and
tight ends, the number or yards
gained as the recipient of a pass from
another player throughout the
subject’s NFL career.
Yards Per Catch For wide receivers and tight ends, the
!
39!
average number of yards gained per
reception throughout the subject’s
NFL career.
Pro Bowls For all offensive linemen, the amount of
Pro Bowl selections received
throughout the subject’s NFL career.
Note: This honor is voted on by NFL
fans and awards 86 players.
All-Pros For offensive linemen, the amount of
Associated Press All-Pro selections
throughout the subject’s FCS or
Division I career.
Note: This honor is voted on by
sportswriters and awards 28 players.
Because of the voting population,
and the significantly fewer amount of
spots, it is considered a higher, and
more rare, honor than Pro Bowl
selections (Smith, 2018).
Tackles For all defensive players, the amount of
total tackles made throughout the
subject’s NFL career.
Tackles For Loss For all defensive players, the amount of
tackles behind the line of scrimmage
made throughout the subject’s NFL
career.
Sacks For all defensive players, the amount of
times the subject tackled the
quarterback behind the line of
scrimmage on a pass play in the
subject’s NFL career.
Forced Fumbles For all defensive players, the amount of
times the subject stripped the ball
from an opposing ball carrier
throughout the subject’s NFL career.
Fumble Recoveries For all defensive players, the amount of
forced fumbles possessed to
complete a turnover throughout the
subject’s NFL career.
!
40!
Interceptions For all defensive players, the amount of
passes caught to force a turnover
throughout the subject’s NFL career.
Passes Defensed For all defensive players, the amount of
pass breakups added to the number
of passes intercepted throughout the
subject’s NFL career.
Combine Drills
The Combine measures will be collected from Pro-Football-Reference.com, as
their database contains Combine drill results on every player that attended the Combine
beginning in 2000. Should Pro-Football-Reference.com lack any needed data,
NFLCombineResults.com will be the secondary source to verify, as the website is
devoted to the single purpose of housing results all the way back to 1987, the year the
different Scouting Combines merged into one. In some cases, a prospect may choose not
to participate in a drill at the Combine, therefore the prospects’ Pro Day, which gives the
prospect a private workout with scouts and a chance to do any drills in which they did not
participate at the Combine, will be used to fill in any holes in this study’s data. This was
deemed a correct course of action because the Combine is simply the vehicle of
conveyance for the drills that scouts most take into account, and if a prospect does not
run, lift, or jump at the Combine, the prospect is still capable of doing the exact same drill
at their Pro Day.
Table 4.
NFL Draft Combine Drill Descriptions
DRILL DESCRIPTION
40-Yard Dash The 40-yard dash is the marquee event at
the combine much like the 100-meter
dash at the Olympics. The drill
!
41!
measures speed, explosion, and burst.
These athletes are timed at 10, 20 and
40-yard intervals. Scouts are looking
for is an explosion from a static start.
Vertical Jump The vertical jump is a test to measure
lower-body explosion and power. The
athlete stands flat-footed and officials
measure his reach. It is important to
accurately measure the reach, because
the differential between the reach and
the flag the athlete touches, placed
directly above where the athletic
begins his jump, is his vertical jump
measurement.
Bench Press The bench press is a test of strength –
225 pounds, as many reps as the
athlete can get. What the NFL scouts
are also looking for is endurance.
Anybody can do a max one time, but
what the bench press tells the pro
scouts is how often the athlete
frequented his college weight room
for the last 3-5 years.
Note: These definitions were adapted from the NFL’s official website describing NFL
Combine workouts (“What goes on”, 2019).
Mode of data Gathering
The data will be collected via secondary data analysis because of the rigorous
collection process that has been proven by the already-existing data sources. In the face
of the impressive tumult through which our online sources, as well as the Combine itself,
go to when ensuring their data is correct, the authors could not have recreated as reliable
of a process as was already in place for either phase.
As an example of the Combine’s strenuous collection, the 40-yard dash, one of
the drills examined in this research, uses a system of lasers set up at the beginning, 10
!
42!
yards, 20 yards, and end of the 40-yard area in which a prospect runs to time their
progress, while stopwatches operated by scouts are used as a backup timing method.
Additionally, the reputability of the Combine gives even more backing to the
stance that using any other kind of method to collect this data and evaluate it would not
be useful. The Combine has 36 years of results behind it that have not been questioned by
pundits or teams selecting the players, so taking it at face validity was determined the
best course of action by the authors.
Secondary data collection of collegiate and NFL statistics, much like with the
NFL Combine numbers, once again came out as the predominant favorite in past studies,
and the authors for the current study will use the same. Since game statistics have been
the way that athletes have been evaluated over the life of sports, choosing to match those
against each other, and against Combine measurements, will be a reliable and
representative way in which to go about this study.
Reasoning for Collegiate & NFL Statistic Choices
The statistical categories measured for the collegiate and NFL portion of data
collecting were used for a variety of reasons, from previous research all the way to
modern-day reclassifications of important statistical categories according to top sources.
Defensive players were all tracked using the same statistics: Games Played,
Tackles, Sacks, Tackles For Loss, Fumbles Forced, Fumbles Recovered, Interceptions,
Passes Defensed. While defensive linemen, linebackers, and defensive backs all have
different duties, by splitting players up into position groups, a median and standard
deviations off that median create a baseline to compare each prospect’s statistics.
Because of this, while some position groups will have greater statistics than others
!
43!
because of the nature of their duties, no danger exists in misidentifying a player’s
strengths or weaknesses based off these stats because they will draw direct comparisons
to those that perform the same duties. These statistical categories were the ones chosen
for defensive prospects because they have long been the most popularized, measurable,
and opinion-determining factors in mass media and sports culture when it comes to
defensive measurements, as discussed earlier in this section via the Alexa ranking and top
sports websites statistical tracking of football players.
Similarly on offense, the same can be said for quarterbacks, running backs, tight
ends, and wide receivers whose stats were taken into consideration. Nearly all of the stats
measured for each position have long been engrained in the collective consciousness of
those that follow football, so the choice was clear on completion percentage (QB),
passing yards (QB), rushing yards (QB, RB), yards per carry (RB), touchdowns (QB,
WR, RB, TE), receptions (WR, TE), receiving yards (WR, TE, RB), and yards per catch
(WR, TE). Receivers, both wide and tight, have been judged on their ability to catch the
ball consistently for eons, hence the receptions statistic. Those that average a high yards
per catch are considered of high value because of their ability to pick up many yards on
one play, while those with high yards and touchdowns show a penchant for reliability and
big plays, respectively. Running backs not only are judged on rushing yards, touchdowns,
and yards per carry, the three most representative popularized statistics of success, but
also receiving yards because the added dimension of being able to catch the ball brings an
extra threat for defenses to base their game plans around. Quarterbacks, much like
running backs with receiving yards, are tracked with rushing yards as well, as the ability
to run while also executing the top duty of a quarterback - throwing the ball for yards,
!
44!
touchdowns, and at a high percentage – brings more to the position than the average
quarterback may present. The one statistic included outside the normal range of measured
success is one involving quarterbacks, yards per attempt, which directly correlates with
winning perhaps more than any other long-examined stat (Barnwell, 2017).
Collection on these “skill” offensive players is the most straight forward, as their
statistics are most easily quantifiable and measured most precisely and thoroughly of all
positions because they are directly responsible for football’s goal - yards gained and
touchdowns.
The only group that presents significant statistical difficulty is offensive line. As
Sports Reference’s Director of Football Operations Mike Kania shared, with their
contribution so hard to quantify because of a lack of knowledge about the blocking
scheme or assignment on each play, statistics could be missing that a player that was
judged to have missed on a block was actually doing the right thing. While sack rates,
pancake blocks, and rushing yardage per side statistics may draw closer to effective, even
those seem to have much dependence on the quarterback or running back’s ability to
recognize what's going on around him and adjust accordingly (Kania, 2016). Previous
studies have excluded offensive linemen because of a statistical dearth, while Lyons et al.
(2011) cited an abundance of negative performance information over positive
performance information as the reason offensive linemen were excluded from their study.
This work did not deem it necessary to exclude offensive lineman since they participated
in the same drills and can be judged in the same baseline ways that other prospects may
be (games played, years played, Pro Bowl appearances, All-Pro honors). While the
information on offensive linemen will not be as in-depth as other prospects, height and
!
45!
weight being measured may help in giving a size predictor towards success in the NFL,
while Combine drills can be cross-examined against games played to extrapolate the
predictor that best determines longevity, which can be correlated with success.
No information was collected on kickers or punters because none were selected in
the first round over the sample period.
Sample Timeframe
The years in which data is collected had to be determined in order to obtain what
was deemed a representative sample of the population. The timeframe of the event in its
current form is 1982-2018, therefore the population is 36 years of NFL Draft first-round
selections. Considering recent advances in the Combine and its evaluators, years 2007-
2012 were chosen, this timeframe selected because the Combine began to be maligned
and popularized in the mid-1990s, became the media event it is today in the late-90s and
early-2000s, and only developed into the spectator event it currently is within the last 15
years. Prior to the mid-90s, the onus and expectation for a top performance at the
Combine was not as prevalent as it is today because the Combine did not impact draft
stock as much as it does now.
The sample, which covers nearly 20 percent of the population of first-round draft
picks since the Combine began, will be large enough to extrapolate results on the entire
population, and with the majority of careers in the sample having already reached their
conclusion, the authors will be able to see how the standard first round draftee’s career
plays out and what indicators within collegiate game performance and Combine
measurables predict that performance accurately.
!
46!
Additionally, the analytics revolution in sports that is now used to judge prospects
and current professionals has made significant progress in recent years (Clark, 2018), so
to capture a population inside that revolution that will be representative of recent and
future results would seem necessary to scrutinize the most recent draftees eligible for the
study.
“Power 5” Conferences v Non-“Power 5” Conferences
Finally, when it comes to codifying conferences of prospects, we turned to the
Lyons et al. (2011) explanation for why this step is needed:
“This variable was controlled for because of the increased probability that high
ability players would be more attracted to competing at a BCS institution and, in
turn, a BCS institution would be more willing to offer these players scholarships
(see Schneider, 1987). Further, players who compete within BCS conferences
may face more competitive teams, which then may attenuate their collegiate
performance—conversely, players of high ability competing at the non-BCS level
may accumulate greater performance statistics than those at the BCS-level due to
lesser competition.” (Lyons et al., 2011, p. 164)
Lyons et al. (2011) dichotomously coded the value of a player being in a Bowl
Championship Series (BCS) conference at the time of his study, simply assigning a 1 to
players within a BCS conference and a 0 to players not within one. As Lyons et al. (2011)
discussed above, the reason for this is because of the difficulty to amass significant
statistics against better competition, as the BCS claimed to have, while the level of
competition is perceived to drop at a prospect’s school that is not within the BCS,
therefore lessening the competitive barriers to success.
!
47!
The BCS, though, was abolished in 2012 amidst controversy that it was not
accurate in determining which teams should make the NCAA postseason and participate
in its most prestigious and significant bowls. Having awarded the national championship
in its most prestigious game to the Southeastern Conference (SEC) six years in a row, the
final of which came about after questionable voting led to Alabama getting in over
Oklahoma State of the Big 12, fans that long-wanted a College Football Playoff that was
meant to level the playing field and stop favoritism and exclusionary practices that
existed in the BCS got their wish (Staples, 2018). This type of playoff system existed at
every level of college football and college basketball aside from the top collegiate
division in the country, FBS Football, since its move to bowl alliances that first formed in
1992 opted for a postseason that penalized non-agreement conferences, creating a
possible competitive imbalance (Eckard, 1998).
With the establishment of the College Football Playoff, while “BCS conference”
is a term no longer used, the increased use of the term “Power 5” has taken over. This
term recognizes schools in either the ACC, Big 12, Big 10, Pacific 12, or Southeastern,
half of the conferences in the FBS (Eckard, 2018). These conferences still represent the
majority of college football’s power, occupying 190 of the 200 Associated Press’ top 10
slots over the last 20 seasons at year’s end. Additionally, current “Power 5” schools,
including Notre Dame, an independent by choice that is in a Power 5 conference in every
other sport, had a record of 87 wins and 15 losses against non-Power 5 FBS schools, a
.853 winning percentage in 2017. The record against non-FBS schools was 39-1, a .975
winning percentage (Eckard, 2018). These statistics establish a competitive imbalance in
favor of the FBS, as well as one within it, “Power 5” schools disposing of those outside
!
48!
of the classification roughly 17 of every 20 times. Further, in 2014, the special status of
“Power 5” conferences was recognized with a grant of significant autonomy in terms of
governance and rule-making within the NCAA’s overarching regulatory structure. Of this
act, ESPN Senior College Football Writer Ivan Maisel (2014) levied the following:
“The NCAA Board of Governors' vote to grant autonomy Thursday to the five
biggest revenue-producing FBS conferences and Notre Dame should be
remembered as a historic day in intercollegiate athletics. On this day, the NCAA
voted that the strong shall inherit the earth.
This means self-rule, and if you hear it and think of the downtrodden rising to
smite their oppressors, then the spin of the Mike Slives of the world has achieved
its goal. Slives, the Southeastern Conference commissioner, said last month, "If
we do not achieve a positive outcome under the existing big tent of Division I, we
will need to consider the establishment of a venue with similar conferences and
institutions where we can enact the desired changes in the best interests of our
student-athletes.
In other words, if you don't grant us autonomy, we will establish autonomy.”
(Maisel, 2014, para. 1)
Indeed, autonomy, and continued power was handed to the five conferences.
Because of this autonomy, the statistics put forward by Eckard (2018), and Lyons
et al.’s (2011) system, this study will measure success of both Power 5 and non-Power 5
prospects, all the way down to the individual school they attended, to see if Power 5 stats
need to be weighted in later studies because of the difficulty that prospects facing the
more impressive competition had to navigate. The former Big East (now AAC) will be
!
49!
included amongst Power 5 schools because it was, at the time the prospects in this study
were in college, a BCS school, therefore we must use the playing field that the prospects
were competing on to impress the overall point that the top-level postseason models, once
the BCS and now the Power 5, either present a poor or strong indicator of NFL success
for players that attend those schools. Additionally, and for clarification purposes, the
school’s conference at the time of the sample, that being 2007-2012, was the one used.
This may not be the conference that a school is currently in, as conference reformation as
a result of the elimination of the BCS and other factors led to programs seeking different
leagues, the case for 19 of the 65 schools that had first-round draft picks. This, though,
will not have an affect on the author’s work, as the school’s ranking average from the
eight-year ranking collection period will be telling and projectable regardless of
conference, giving a definitive answer on the subject of if facing tougher competition in
college is a successful determinant of future NFL success.
The process for determining which schools, and therefore conferences, were the
strongest at the time of the sample was important to ensure schools were being ranked
appropriately. To do this, the authors turn to the widely-respected Sagarin Rankings,
which represent the average schedule difficulty faced by each team in the games that it's
played to a given point in the season, the schedule difficulty of a game taking into
account the rating of the opponent and the location of the game (Sagarin, 2019). The
rankings are a combination of two computer-based point systems that are used to
generate a final rating, and also log college football teams’ record against top-10 and top-
30 teams in the rankings. For nearly 50 years, these rankings have been part of the
!
50!
mainstream consciousness and one of the most trusted ways to rank college football
teams (Feng, 2019).
Founded by Jeff Sagarin during his time at MIT, the right to publish his schedule
ratings was purchased by Pro Football Weekly in 1972, began to be used by college
basketball’s NCAA tournament in 1984, have been part of USA Today’s college football
coverage since 1985, and were one of just three original computer rankings used by the
BCS in its inaugural year of 1998 (Niesen, 2018). Despite the BCS’ downfall, Sagarin
still has his rankings published by USA Today as of the conclusion of this study.
After the Sagarin Rankings were selected and reviewed for legitimacy, the question
of timeframe measured arose. The rankings measure the college football season prior to
the NFL Draft is conducted every season, so with the sample consisting of NFL first-
round draft picks from 2007-2012, it became obvious that the Sagarin Rankings that were
included must stop at 2011, as the 2012 rankings would be measuring the 2012 college
football season that usually begins in August, the draft taking place roughly four months
prior on an annual basis. Further, if this study was to capture sufficient data on players
that were being selected in the first year of the sample, that being 2007, we could not
simply take the 2006 Sagarin Rankings and apply them as holistically representative to
the first year of our sample that was drafted in 2007. Rather, it would be necessary to
travel back a couple of years before the sample began because the players selected in
2007 were playing significantly at their collegiate institutions beginning in either 2004 or
2005, with the rare exception amongst the sample seeing a freshman or redshirt freshman
in the 2007 draft class stepping on the field in 2003. Thus, the period of Sagarin Rankings
collected ran from 2004 through 2011. In collecting this data, the final rankings from
!
51!
each season were the ones acquired, logged, and will be averaged to determine where
each school was rated in relation to 238 others that the rankings cover over the eight-year
period.
As an example, Louisiana State University, a member of the Southeastern
Conference, which has been considered the top league in the nation eight of the last 13
years by the rankings (Sagarin, 2019) and has produced 10 national champions in the last
13 years, ranked numbers 18, 7, 3, 1, 21, 13, 8, and 2 in chronological order from 2004-
2011 in the final poll of each year collected for this portion of the study. This produced
an average Sagarin Ranking of 9.125 over the eight-year period studied, one of the best
measured during that time. The process of data collection, logging, and averaging will be
repeated for every other school that had a draft pick during our sample’s 2007-2012
period, and studied based on conference and individual institution.
!
52!
CHAPTER 4
RESULTS
Because of the different expectancies of different positions in the NFL, and the
like-comparisons we can make between different prospects at the same position, the data
was broken up by position, averages taken of the individual statistics amongst the groups,
and results of individual prospects measured against the averages. There were 17
quarterbacks, 17 running backs, 21 wide receivers, four tight ends, 36 offensive linemen,
21 defensive tackles, 26 defensive ends, 19 linebackers, 22 cornerbacks, and 10 safeties
in the six-year collection period. 86 of the athletes were still active in the NFL as of the
collection data, which was just prior to the 2018 regular season, while 105 athletes had
concluded their careers. Because the majority of the careers for the active players had
concluded at the time of the collection process (based on the average of 9.3 years that a
first-round pick lasts in the NFL), the fact that players were still active was not a concern
in this study, as at least six seasons had elapsed for every player in the sample, enough
time to determine a player’s career trajectory. Notable in the sample outside of positional
analysis of success based off combine and college performance were as follows:
-Only 15 of the 191 players were drafted during the sample period were selected from
schools outside the aforementioned “Power 5” & Big East conferences. From this group
came booming success stories such as Joe Flacco, Chris Johnson, Joe Staley, Muhammad
Wilkerson, and Dominique Rodgers-Cromartie, but also some of the poorest performing
prospects in the sample period such as Shea McClellin and Larry English.
- Despite there being significantly more offensive linemen on the field at any given time
than any other position in football (always a minimum of five, no other position has a
!
53!
traditional minimum of more than three), there were not a lopsided number of offensive
lineman drafted in the first round as compared to other positions. While 36 was the most
of any position, it was only 10 more than defensive ends (usually two on the field) and
only 19 more than running backs or quarterbacks (almost always one on the field). While
this could tell many things about the significance teams deem other positions to have over
the offensive line, for the purposes of this study, it highlights, and perhaps correlates with
NFL franchise decisions, that this is the most difficult position to measure. Statistics for
offensive linemen are not kept publicly outside of the amount of games a lineman plays.
Because of the collective nature and scheme of blocking, offensive line contribution is
hard to quantify, dependent on play call, and less obvious in success comparatively to
other positions (Kinia, 2016). This could lead to an uneasiness in spending significant
amounts on linemen early in the draft, but more likely based on research done by SB
Nation’s Arrowhead Pride (What The Stats, 2015) lend credence to the theory that a
quality lineman can be found throughout the draft.
-With six years of the 37 since the NFL Draft Combine came into existence covered, it
seems apparent that our sample will be representative of the population, though two
positions stand out that require further clarification. There were only 10 safeties drafted
in the sample period, none selected in the first round in 2009 or 2011. Historically,
though, this aligns with general draft practices, as only 45 safeties since 1982 have been
drafted, covering roughly 18 percent of the population. Similarly, tight ends were few and
far between amongst first round picks from 2007-2012, only Greg Olsen, Dustin Keller,
Brandon Pettigrew, and Jermaine Gresham chosen with the first 32 picks those years.
While 35 tight ends have been chosen since the Combine’s inception, the sample still
!
54!
represents 11 percent of the population, which admittedly leaves some level of desired
assurance unmet, but will still serve as a reasonable representation for this study.
Statistics for both collegiate and NFL prospects, as well as Combine performance,
were verified by a third party, giving this study inter-rater reliability and leading to a high
level of confidence in the data throughout this work.
Positional Analysis
Quarterbacks. In order to determine which is a better predictor of NFL success,
collegiate performance or the NFL Combine, we must be sure to define success. Our data
can do this, but deciphering which quarterbacks have had the best NFL careers in our
sample can also be assisted by expert opinion. Bill Bender (2018) of The Sporting News
compiled a list of first-round quarterback selections since the year 2000, ranking them 1-
48, and while this is far from exhaustive or conclusive of expert opinion, paired with the
numbers collected from the sample period, it gives solid backing to preconceived
statistical notions. Four from the sample period made the top 10 of Bender’s list – Matt
Ryan, Cam Newton, Joe Flacco, and Matthew Stafford - while six of the bottom 10 were
in the sample – Jamarcus Russell, Brady Quinn, Brandon Weeden, Tim Tebow, Christian
Ponder, and Jake Locker. The NFL numbers tell the same story the rankings do, with
each of the six quarterbacks listed in the bottom ten lasting four years or less in the NFL,
all at least five years less than the average first round draft pick, while the four in the top
10 of Bender’s list are still active and in the top four statistically from the sample in
passing yards, touchdowns, and games played. Within the bottom six quarterbacks, the
high, and low, marks for collegiate completion percentage and rushing yards reside,
while the top games played, passing yards and touchdowns all came from the six that
!
55!
proved poor selections in the first round. Of the four that ranked in Bender’s top 10, and
were the best statistical NFL quarterbacks in our sample, the low mark for games played,
passing yards, and touchdowns were found, while the high for yards per attempt was
delivered by the same man responsible for those lows, that being Newton. The low for
yards per attempt came from Locker, meaning the congruence between NFL success and
collegiate game performance existed, at least in the extremes, most often with yards per
attempt.
Evaluating Combine performance is more difficult with quarterbacks than other
positions, as 15 of the 17 in this work’s sample chose not to do the bench press, Weeden
and Quinn the only exceptions, while five did not participate in the vertical jump. This is
because a non-quantifiable element of pre-draft workouts, the quarterback’s Pro Day and
private workout, tend to mean more, and are where most of the evaluation of a
quarterback is done (Gabriel, 2018). Based on the two drills the majority of quarterbacks
did, the most prolific in them was Robert Griffin III, running the fastest 40-yard dash and
leaping 39 inches, the highest amongst the 12 participating quarterbacks in the vertical
jump. Griffin’s marginal success put him 28
th
on Bender’s list, and while he is still active
in the NFL, has served as a backup in his most recent seasons. The slowest 40-yard dash
times included Mark Sanchez, largely a backup in his NFL career, Ryan, who ranks in the
top 10 in career passing yards in NFL history, Weeden, who played in just 35 games in
his career, Josh Freeman, who does not stand out for the good or the bad amongst the
sample, and Flacco, who finished last in vertical jump but has a Super Bowl victory, and
a career that rivals Ryan’s, to his name.
!
56!
84 correlations were examined between the Combine as well as collegiate game
performance and their predictive validity towards NFL success. Pearson’s correlation
coefficient was the equation used to study the relationship, or lack thereof, between the
Combine and NFL performance as well as collegiate game play and NFL performance.
Of the 84 possible correlations, only six showed significance. Of those six, four were
negative, indicating that when one categorical statistic increased, the other being
measured in the correlation equation decreased. For quarterbacks, this was the Combine
drill vertical jump, a negative relationship significant at .001 with NFL games played,
NFL years played, and passing yards, a negative relationship significant at .005 with
touchdowns.
The negative correlation between vertical jump and NFL success in four of the
seven categories measured in this study implies that athleticism does not correlate to NFL
success at the quarterback position. Mcgee and Burkett (2003) showed interrelationships
between the vertical jump and the sprint drills as well as the broad jump when testing
variables of their study for correlation. This is to say, if a player performs well in the
vertical jump, they should also perform well in the 10-, 20-, and 40-yard dash drills.
Indeed, six of the top seven quarterbacks according to vertical jump in this study’s
sample also possessed the best times in the 40-yard dash. These quarterbacks would be
considered “athletic” when placed against their peers, but four of the seven ranked in the
bottom third in Bender’s ranking of first-round quarterbacks listed above. Three of those
quarterbacks are no longer in the NFL, all lasting four years or less, less than half the
shelf life of the average first-round draft pick.
!
57!
These findings further support the group of NFL enthusiasts that claim a team
cannot achieve the ultimate goal, winning the Super Bowl, with a “mobile” or “dual-
threat” quarterback. This study has just one Super Bowl winner in it, Joe Flacco, whose
vertical jump was worst amongst the sample group and whose 40-yard dash was in the
bottom third. Those that were successful in this sample’s drills have rarely been
successful in the NFL, with the lone exception being Cam Newton, who has made one
Super Bowl, losing to veteran pocket passer Peyton Manning. While Newton has come
close, only two Super Bowl winning quarterbacks in the last 20 years have rushed for
more than 225 yards in a season, a negligible 15 yards per game, during their careers,
those being Russell Wilson and Aaron Rodgers.
Quarterbacks rarely have to jump, but they often have to run, and with how the
vertical is associated with running and athleticism, this study strikes another blow to the
“dual-threat” quarterback, showing that the ability to pick up yards with your feet pales in
comparison to the ability to consistently pick up big yards through the passing game.
Running Backs. The list of first round running backs is exactly as long as that of
quarterbacks in the author’s sample, and unlike the previous position, there are more
successful cases once the prospects reached the NFL than underwhelming ones.
NFL.com’s Gregg Rosenthal (2012) listed the best first round running backs of the
previous decade with four of the top six coming from the sample. Additionally, three of
the running backs – Marshawn Lynch, Ryan Mathews, and Mark Ingram - that did not
make the top six only improved their cases over the years with healthy and extended
careers. Three of the sample’s running backs finished in the bottom eight, comparatively
small to this study’s quarterbacks placing six in the bottom 10 of Bender’s (2018) list.
!
58!
The clear cut, top-performing backs were Adrian Peterson, No. 1 on Rosenthal’s
list, Marshawn Lynch, who, along with Peterson, is one of just five amongst the running
backs that is still playing, and Chris Johnson, who, also with Peterson, is one of seven
men in NFL history to rush for over 2,000 yards in a single season. Peterson led the
sample in touchdowns and yards, Lynch coming in second in both rankings, while
Johnson led in receiving yards and was third in rushing yards and touchdowns. Also
having success were Ingram, Jonathan Stewart, and Darren McFadden. Contrarily, David
Wilson lasted only two seasons in the NFL, collecting just 504 yards, narrowly trailing
Jahvid Best, who found more success on Olympic sprinting tracks than on the football
field (Kirshner, 2016). Trent Richardson, Beanie Wells, and Donald Brown round out
others that did not produce to their draft position.
Collegiately, it seemed apparent that Peterson and McFadden would have success,
marking the top two in rushing yards and accounting for two of seven in the sample to
finish with more than 40 collegiate touchdowns. At the bottom of rushing yards
production were Wilson, Best, Knowshown Moreno (who played the least college games
of any in the sample), and Rashard Mendenhall, three of which finished in the bottom 10
of Rosenthal’s running back list and all of which finished below the average yards gained
in the NFL amongst the sample. Wells finished with the least collegiate receiving yards
while CJ Spiller gained the most (while playing in the most games), Doug Martin topping
touchdowns with Felix Jones, the collegiate yards per carry leader, coming in last in
scores. While Brown’s 5.4 yards per carry, the least amongst the 17 running backs, gives
a hint at that stat projecting success well, Jones’ top mark of 7.7 eschews that theory,
!
59!
though rushing yards and touchdowns, with top collegiate numbers coming from the most
productive NFL backs, generated encouraging feedback.
From the Combine, running backs represent one of the more athletic positional
groups amongst the 10 measured, logging the fastest individual 40-yard dash time,
second-highest individual vertical jump, and quality performances on the bench press.
The 40-yard dash time came from Johnson, whose 4.24 was fastest in Combine history
until 2017 when receiver John Ross ran a 4.22. Johnson’s speed was the trademark of his
career and contributed to his solid NFL career, though Ingram, who finished last in the
drill amongst first round selections in the sample with a 4.62, has not seen his slower time
hinder him in the NFL. Peterson and Lynch, the top producers amongst running backs,
both finished middle of the pack of the 17 participants, signaling little connection
between the drill and NFL success. Results from the bench press were mixed, with long-
time NFL impressers Martin and Stewart tying for the group lead with 28, but McFadden
and Jones, whose NFL success was mixed, coming up with a tie for least reps at 13.
Initial reviews of vertical leap seem to suggest an inverse relationship, with Brown and
Wilson marking the top two numbers by a significant margin, and Ingram, despite his
31.5-inch jump, lowest out of the backs, continuing his NFL career to this day.
66 correlations were examined between the Combine as well as collegiate game
performance and their predictive validity towards NFL success. Pearson’s correlation
coefficient was the equation used to study the relationship, or lack thereof, between the
Combine and NFL performance as well as collegiate game play and NFL performance.
Of the 66 possible correlations, only one showed significance. That one correlation
showed a positive relationship, significant at .005, between collegiate rushing yards and
!
60!
NFL receiving yards, suggesting that a productive running back in college running the
ball, even if they may not have large receiving numbers collegiately, have the ability to
produce in a variety of ways in the NFL. This was seen with Peterson, Lynch, and
Mcfadden, who were in the bottom half in collegiate receiving yards, but were second,
third, and fourth in NFL receiving yards, productive enough to affect the game in a
variety of ways.
These findings contradict those of Kuzmits and Adams (2008), who generally had
similar findings to this study, with the exception of sprint times maintaining correlational
significance to running backs NFL success. While Chris Johnson’s 4.24 stands out as
spectacular against, not only the sample, but the history of the Combine at any position,
Ingram’s 4.62 served as no barrier to having success of his own. Mcfadden’s success was
similar to Ingram’s, though his 4.33 time was second-best amongst the sample, while four
running backs ran a 4.46, Lynch and Stewart excelling in the NFL, Martin and Brown the
opposite. These numbers do, however, support Robbins (2010) and Park (2016) in the
finding that 40-yard dash time most influences NFL draft position amongst running
backs. A study done by SB Nation’s Mile High Report (Doll, 2013) of 40-yard dash
times from NFL draftees between the years of 2000 and 2012 showed the average
running back’s 40-yard dash time was 4.49. Amongst the 17 running backs drafted in the
first round between the years of 2007 and 2012 in this work, 15 of them ran faster than
that time, nearly half coming in a full tenth of a second below the average mark on a
scale in the sample that was less than four tenths wide.
Wide Receivers. With 21 wideouts in the sample, many of whom went on to
tremendous success and continue their NFL careers today, wide receiver is a strong
!
61!
position in the sample, and one of the three positions Mcgee & Burkett (2003) believed
would have a strong correlation between Combine success and draft success. Along with
running back and defensive back, the researchers found in their work that vertical jump,
10-yard dash (found to be highly correlated with the 40-yard dash measured in this
study), and three-cone drill (not measured in this study) had a direct correlation with
which round the prospects at the above positions would be selected during. While this
differs from this study’s ultimate goal of having the Combine predict NFL success, the
connection is apparent that expected success is correlated with draft position, therefore
forming a level of symmetry between the studies. Mcgee and Burkett stated that the three
positions were most often affected by these drills in their draft position because they are
the spots on a football field that most rely on explosive speed and agility. Their work
guides this study to examine Combine numbers even closer, and one example that proves
their theory is Darrius Heyward-Bey, who had the fastest 40-yard dash time amongst the
21 receivers in the sample at 4.3, leading him to be selected seventh overall in 2009
despite never breaking 800 yards receiving in a season at the University of Maryland. At
the low end of the 40-yard dash time was Michael Crabtree, who led the receivers in
collegiate touchdowns and remains a consistent presence in the NFL today despite his
subpar run that saw him drafted 3 spots behind Heyward-Bay despite a far better college
resume. Leading the receivers in bench press was Kenny Britt, putting 225 pounds up 23
times in his workout, while Kendall Wright was able to do so just four times to mark the
low number for wideouts. The most successful professional receiver in the group, Calvin
Johnson, leapt 42.5 inches, a half-inch clear of Jonathan Baldwin for tops at the position,
with Dwayne Bowe coming up 9.5 inches short of Johnson to mark the low end.
!
62!
In their collegiate careers, on the opposite end of Crabtree’s 41 touchdowns (in
just two seasons) was Craig Davis with seven, Davis lasting only four years in the NFL,
one year ahead of Jonathan Baldwin, AJ Jenkins, and Justin Blackmon for fewest
amongst receivers. Wright, despite his meager performance on bench press, led the
position in games played, receptions, and receiving yards in college, his NFL career later
ending after six seasons of up-and-down production. The low mark for receptions and
receiving yards was Anthony Gonzalez, whose little collegiate production and average
Combine numbers translated into a quiet professional career in which he played just 40
games. In yards per catch, the runaway leader was Demaryius Thomas, whose nearly 20
yards per catch hoped to offset a poor 40-yard dash of 4.52, while Jeremy Maclin, known
for his NFL production down the field, only averaged 12.7 yards per catch, lowest in
collegiate stats out of the receivers.
66 correlations were examined between the Combine as well as collegiate game
performance and their predictive validity towards NFL success. Pearson’s correlation
coefficient was the equation used to study the relationship, or lack thereof, between the
Combine and NFL performance as well as collegiate game play and NFL performance.
Of the 66 possible correlations, only four showed significance. Those four all stemmed
from body measurements, height of the receivers positively correlating at .001 to NFL
yards per catch, while weight positively correlated at .005 to NFL receiving yards,
touchdowns, and yards per catch. Interestingly, there was only one other positive
correlation to a body measurement throughout the sample, defensive ends weight
positively correlating to NFL games played.
!
63!
The results suggest that size and strength are especially important at the wide
receiver position, while the findings of Mulholland and Jensen (2016), which supported
speed as mildly predictive of NFL success across their longitudinal study of receivers,
seem secondary to the findings of this work. Dez Bryant and Demaryius Thomas, top-
five in NFL receiving yards amongst the sample, each ran 4.52 40-yard dash times,
second-worst amongst the 21 receivers, but weighed in within a pound, and an inch of
each other, both roughly 225 pounds and 6’3”. The sample’s largest receiver was the
aforementioned Calvin Johnson, who is the single-season record holder for receiving
yards in NFL history and leads the sample in touchdowns, receptions, and receiving
yards, weighed in at 239 pounds and measured 6’5”, both sample-highs. Additionally, the
other two in the top-five in NFL receiving yards amongst the sample were AJ Green,
6’4”, and Julio Jones, 6’3”, 220 pounds. Bryant, Thomas, Johnson, Green, and Jones
were all above the average of 6’1.5”, Johnson, Green, Thomas, and Jones measuring as
four of the top five receivers. Additionally, with the average receiver in the sample
weighing 210.5 pounds, Johnson outweighed the sample average by nearly 30 pounds,
Thomas and Bryant by roughly 15 pounds, Jones by nearly 10.
The need for top-level size amongst receivers stems from the fact that they are,
for the most part, matched up on cornerbacks. In this sample, corners average weight was
196 pounds, while their average height stood 5’11.5”. Two inches and 15 pounds may
seem miniscule, former Denver Broncos Head Coach John Fox summarized the simple
premise behind the advantage wide receivers now have in the modern passing league the
NFL has become: “This league is a bigger, faster, stronger league, and you win when you
win matchups. If you're the bigger, faster and stronger guy, you're going to win more
!
64!
matchups. That's not rocket science there. That's just the way it is.” (Legwold, 2014, para.
4). While speed did not predict NFL success in the current study, receivers have the
advantage there too, 21 running faster than a 4.49 40-yard dash, only 12 cornerbacks
doing so at the 2014 Combine (Legwold, 2014). The combination of height, weight, and
speed makes the point apparent that wide receivers that boast the size of those that have
been tremendously successful in this work’s sample hold a dynamic advantage, and have
changed the game since the size differential was non-existent. In 1992, Pro Bowlers at the
cornerback position and the wide receiver position were nearly exactly the same size,
with cornerbacks even outweighing receivers by six pounds. One year after this work’s
sample was complete, 2013, receivers in the Pro Bowl were three inches and nearly 20
pounds heavier than their counterparts on the defensive side of the ball (Legwold, 2014).
A direct physiological advantage for the five success stories of this position, it’s
no surprise to see the six receivers that weighed in at less than 200 pounds average just
over 500 yards per season, roughly 200 yards per year less than the average receiver in
the sample.
Tight Ends. The smallest group within the sample, Greg Olson, Dustin Keller,
Brandon Pettigrew, and Jermaine Gresham are the four tight ends drafted in the first
round from 2007-2012. This position was studied by Mulholland (2014) in an attempt to
better understand the changing nature of the position in a league that is using it more to
pass than as a blocker as contrary to years past. The study found that the only predictors
(amongst over 30 variables between both collegiate performance and Combine results
that were measured) of both draft order and NFL success were weather a player attended
school at a BCS (Power 5) school and the 40-yard dash. Predictive of NFL success only
!
65!
was the broad jump, not measured in this study, and predictive of draft order was the
bench press, but nothing aside from 40-yard dash time and BCS indicator were indicative
of both a high draft position and NFL achievements. The reasoning put forth by
Mulholland for the 40-yard dash matched exactly with the goal of the paper – to show
that the evolution of the tight end as more of a pass catcher than a blocker was being
reflected in tight end skill set and the way in which teams were drafting. Since it is
important to possess speed in more of a pass-catching position than a run-blocking role,
as tight end has shifted from, the results in the work matched the hypothesis and the NFL
trend.
Those results, though, were mixed in this study’s sample, with Greg Olsen
representing the most productive NFL pro by a significant margin. Olsen ran a 4.51 40-
yard dash, the fastest amongst tight ends, and was approaching 8000 receiving yards in
his professional career at the time of this study, a number still growing as his career
pushed into its 12
th
season. But while Olson’s 40-yard dash did tell the story of a long
and prolific NFL career, Dustin Keller, who ran a 4.53, narrowly behind Olsen, had the
opposite result. Keller finished last in all but one category of NFL performance
throughout the four tight ends measured, and had only a five-year career. Jermaine
Gresham and Brandon Pettigrew, drafted in nearly identical locations in the first round
exactly one year apart, made up the middle of the four tight ends in regards to pro
production, their 40 times falling short by a sizeable margin to Keller and Olsen.
To believe that collegiate performance would help clear up any confusion on just
how Keller and Olsen were physically similar in Combine performance yet so drastically
different in their NFL careers would be a wise thought, but ultimately an unscrupulous
!
66!
one. Keller was actually the most productive collegiate tight end of the four, leading the
group in receptions and receiving yards, Olsen coming in last in games played,
touchdowns, receptions, and receiving yards. Gresham had the most touchdowns and the
highest yards per catch, while Pettigrew played in the most games collegiately.
66 correlations were examined between the Combine as well as collegiate game
performance and their predictive validity towards NFL success. Pearson’s correlation
coefficient was the equation used to study the relationship, or lack thereof, between the
Combine and NFL performance as well as collegiate game play and NFL performance.
Of the 66 possible correlations, only five showed significance. All five of those were
negative correlations, college games played significant at .001 to pro receiving
touchdowns, also significant at .005 to pro games played, pro receptions, and pro
receiving touchdowns. College receptions also showed a negative relationship with pro
years played, statistically significant at .005.
These results are not a surprise considering Olsen was thoroughly and completely
superior in the statistics measured in this study compared to the other three tight ends,
and also played anywhere between eight and 14 fewer games collegiately than his
position-mates in the sample.
This position, despite having a representative sample of the population, over 11
percent included in this study, needs further investigation to verify the findings because
the sample does not have proportionality to the amount of tight ends in the NFL, and
because of the continuing evolution of the position in a largely pass-heavy league.
Offensive Linemen. An often undervalued, overlooked, and forgotten about
position in the public eye, the offensive line, as discussed above, present a number of
!
67!
challenges in quantifying performance. This has led previous studies, most notably and
recently Vincent, Blissmer, and Hatfield’s (2018) work on this subject, to leave offensive
linemen out all together. Mulholland (2014), in studying a position closely related to
offensive line, studied height, weight, and BMI on top of performance measures at the
Combine. While the author of this work has not discussed the significance of studying the
physical measures of height and weight prior to this, in the conversation of offensive
linemen, it is tantamount to this paper. This is true not only because measures of
offensive line performance are sparse, but also because of the literal physical evolution of
the position in the last five decades. In 1972, the average height of an offensive lineman
was 6’4”, the average weight was 249.6 pounds, and the body fat percentage was only
slightly above the national average at 15.5 percent. In the most recent measure, height
stayed consistent, but the average weight was up 61lbs to 310.6, while body fat
percentage had nearly doubled, measuring 28.8 (Anding & Oliver, 2015). During the
peak of growth amongst linemen, and players in general across different levels of
football, Secora, Latin, Berg, and Noble (2004) found that player growth was
contributing to a power and speed increase as well, with changes in power and speed
drills coming with a statistically significant correlation to physical measurements 50 of
88 times in the study. With those positives for the athletes, though, come possible
negative consequences, as Mathews & Wagner (2010) found that a large percentage of
college football offensive linemen exceeded the at-risk designation of obesity. Given that
offensive linemen are the largest specimens on the field, these size changes, along with
the positive and negative effects stemming from them, are especially important for the 36
linemen in this study.
!
68!
The largest in terms of height was Nate Solder at 6’8”, five including Joe Thomas
and Jake Long right behind Solder at 6’7”. Ben Grubbs and Danny Watkins at 6’3” were
the shortest offensive linemen. This is telling, as Solder, Long, and Thomas all had
notable careers, while Watkins lasted just three years in the NFL. Mike Iupati and Andre
Smith weighed in at over 330lbs, the heaviest amongst the 36 linemen, and continue
playing to this day, though the lightest linemen, Mike and Maurkice Pouncey along with
Joe Staley, have also had tremendous success in the NFL. In addition to years played, to
measure NFL success, this study looks at Pro Bowl and All-Pro appearances for linemen,
as those honors are considered the highest postseason accolades for each position
(Rosenthal, 2013), but will specifically be focused on for lineman due to lack of other
statistical measurement options.
Staley ran the fastest 40-yard dash of any lineman, 4.79, with six-time Pro Bowler
Trent Williams right behind him at 4.81, the slowest runs emerging from Jeff Otah and
Sam Baker, Otah also having the shortest vertical jump. Otah and Baker played just 10
combined seasons with no Pro Bowls made. Strength is synonymous with the offensive
line position, and this study’s measurable for it, the bench press, saw Long, a one-time
All-Pro and four-time Pro Bowler, bench 225 pounds 37 times, setting the pace for the
linemen. Smith was on the other end of the spectrum, doing so only 19 times. As for
collegiate game performance, the only measurable kept was games and years played, and
Otah, James Carpenter, and Watkins all played just 2 years, having marginal-at-best
success in the NFL, while the majority of prospects played all four years. Leading in
games played was Mike Pouncey, four-time Pro Bowler Duane Brown trailing by only
two games, while Otah and Watkins played in just 25 games. Amongst the group,
!
69!
Thomas led in Pro Bowl selections, All-Pro honors, and games played in the NFL, while
Derek Sherrod, Otah, and Watkins were by far the bottom in games played and received
none of these honors. Sherrod was a four-year collegiate with slightly below average
Combine numbers.
Despite a lack of in-game data to measure, offensive linemen had the highest
percentage of statistically significant correlations in relation to correlations possible
amongst the sample. With only 28 correlations calculated, nine came back with positive
returns, six of which were significant at .001, the other three significant at .005. Most
notably, the results suggest that being athletic, rather than large, is the most important
factor to NFL success. 40-yard dash was significant at .001 to NFL years played, NFL
games played, and Pro Bowls, while vertical jump was significant at .005 in predicting
NFL games played and Pro Bowls. The other positive correlations revealed the longer a
player remained in college the more ready they were to excel in the NFL, college games
and years played predicting longer NFL games and years played as well.
Athletically, while offensive linemen have appeared to grow to unhealthy sizes
and body fat percentages (Anding & Oliver, 2015), they have also turned into modern
day elite-level athletes (Shook, 2017) that need to be capable of being mobile in the run
game while also stopping much quicker defensive linemen and linebackers coming
towards their quarterback in the pass game. On the ground, there are three popular
running schemes in the NFL (Shook, 2017). The zone-blocking scheme requires linemen
to work upfield rather than simply power block whoever is in front of them, while also
being mobile enough to get to the sidelines if necessary (Bullock, 2015). This, at times,
means linemen will be blocking multiple players of different sizes and at different levels
!
70!
of the field during the same play. In so doing, while they will generally never run 40
yards, their ability to move quickly from level-to-level of the defense and keep up with
the play is vitally important to the fastest-rising run-blocking scheme in the NFL (Shook,
2017). The power running game, while it sounds figuratively, and literally, straight
forward, in fact features guards “pulling” to create deception across the defensive line.
This means that the biggest players on the field are on the move from one spot along the
line to another once the ball is snapped, needing to cover upwards of 20 yards when
doing so, meaning straight line speed is pertinent to ensure success. Finally, the counter
blocking scheme not only brings guards from their original position to pull, but does so
with a tackle too, creating multiple moving parts that requires exceptional timing and
quickness by the linemen up front to meet their responsibilities before the ballcarrier gets
to the hole he needs to run through (Shook, 2017).
In the passing game, linemen must face opposing defenses that blitz defensive
ends that averaged a 40-yard dash time nearly a half-second faster than that of their
position group in this work’s sample, defensive tackles that averaged roughly the same
weight and a quickness advantage of nearly a quarter of a second, linebackers that were
over a half-second faster than the linemen, and occasional defensive backs that outpace
the linemen by even more. All this while the opposing players are free to get a running
start, building momentum to further their force against the offensive linemen, who are
stationary until snap. Slower offensive linemen can struggle to keep up with the amount
of defenders rushing the passer, the schemes designed to confuse and misdirect the
protectors of the quarterback, and simply be beaten one-on-one by a more agile athlete,
where those performing well in the athletic drills at the Combine do not have that issue.
!
71!
Much like quarterbacks, jumping is not something an offensive lineman will often
be asked to do, and running 40 yards in a straight line falls into the same category, but as
demonstrators of athleticism, they give a look into what makes a behemoth along the
offensive line successful, or the opposite. There were six linemen that ran below five
seconds in the 40-yard dash, accounting for 28 Pro Bowl appearances of the 68 the
sample accrued, a disproportionate amount leaving the other 30 offensive linemen in the
sample with just 40 Pro Bowl appearances combined. Additionally, Williams, Thomas,
Staley, and Solder all ranked in the top-seven of the sample. Based off this study, a strong
baseline for selecting an offensive lineman in the first round would be a prospect that has
played four years of FCS or FBS college football, leapt 30-plus inches in the vertical, and
run a sub-five second 40-yard dash at the Combine.
Defensive Tackles. Those closest in size to offensive linemen are the defensive
tackles across from them, this group of 21 weighing in between Nick Fairley’s 291
pounds and Dontari Poe’s 346. The reason for this 57-pound range, the largest variance
amongst any position group in this study aside from wide receiver, which stands at 61lbs,
is because of the prevalent defensive scheme difference in today’s NFL. All teams in the
league line up in different defensive formations, but all will have a “base” package,
which either possesses three defensive linemen or four, labeled the 3-4 or the 4-3. In the
3-4, the larger, space-occupying defensive tackles are selected because they need to
occupy blockers in order to free up linebackers, the level behind the defensive line, to
make plays. In the 4-3, smaller, more fleet-footed defensive linemen like Fairley are
taken, as they are expected to free themselves and make more of the plays than a 3-4
lineman would be expected to make (Tanier, 2005). The reason for the wide receiver
!
72!
range being even larger is the varying positions demanded from receivers, with larger
wideouts playing the furthest away from the quarterback while the quicker, faster, and
generally slimmer receivers playing closer to the quarterback in the “slot”. The largest
receiver in the group was Calvin Johnson, a dominant outside receiver weighing 239lbs,
very large for a receiver, Ted Ginn Jr., a predominantly slot-style receiver weighing just
178lbs., exceptionally small for a receiver. This case could be considered an outlier
because Ginn and Johnson represent significant departures on both ends of usual receiver
weight, defensive tackles often having the largest variance in different samples. Because
of the different schemes and roles of defensive tackles, size is not expected to play a
significant factor in determining success amongst the interior linemen.
Amongst college performers, Ndamukong Suh, the highest pick of any defensive
tackle in the draft during the sample period after being selected second overall by Detroit
in 2010, led six of the eight performance categories, logging the most games played,
tackles, sacks, tackles for loss, interceptions, and passes defensed. The other categories -
forced fumbles and fumbles recovered - were led by Ziggy Hood and Sedrick Ellis,
respectively. On the low end, Fairley and Michael Brockers played in just 27 games,
Marcell Dareus tackled just 71 ball carriers, Brockers accumulated only two sacks, Justin
Harrell can claim only one tackle for a loss, while a bevy of prospects tied for fewest
fumbles recovered, fumbles forced, interceptions, and passes defensed with zero. The
most impressive pro careers in the sample came from Suh and Muhammad Wilkerson,
both generating over 400 tackles and 40 sacks in careers that are still active.
At the Combine, Suh led prospects in the vertical jump, though Suh’s run was
roughly average amongst the sample. The lowest vertical jump was Glenn Dorsey’s at
!
73!
just 25.5 inches compared to Suh’s 35.5, while the massive Poe led prospects in the
bench press with 44 repetitions, sixth-most in draft history (Kaylor, 2017), more than
doubling Brockers’ low mark of 21. Brockers also ran the slowest 40-yard dash, just
ahead of Ellis and Kentwan Balmer, who stuck in the NFL for only nine combined years.
Fletcher Cox, whose year-by-year NFL numbers are amongst the best in the sample, ran a
4.77 40-yard dash to lead all participants.
As was the case with every position on the defensive side of the ball, 126
correlations were examined between the Combine as well as collegiate game
performance and their predictive validity towards NFL success. Pearson’s correlation
coefficient was the equation used to study the relationship, or lack thereof, between the
Combine and NFL performance as well as collegiate game play and NFL performance.
Of the 126 possible correlations, 18, the most amongst any position, showed positive
significance, while only one showed a significant negative relationship. Of the 18
relationships that showed positive significance, 17 came from collegiate game
performance data while only one of a possible 45 Combine correlations came back
positive.
Notable amongst the positive collegiate in-game significance is that neither years
nor games played correlated positively with NFL success. Rather, the five collegiate
categories that were predictive of NFL prosperity were college tackles, college sacks,
college tackles for loss, and college passes defensed. This showed that simply playing a
significant amount in college did not prepare a defensive tackle to be successful at the
next level, as the most successful defensive tackles varied from two to four years of
collegiate experience. Simply being on the field was not enough, though producing once
!
74!
on it showed that college success brings defensive tackles to the NFL ready to make an
impact. Also, it should be noted that NFL front offices became more adept, at least during
this study’s sample, at drafting defensive tackles in 2010, 75 percent of them ranking
above the average for career tackles amongst the sample, the same amount doing so in
sacks, while the 2007-2009 first-round defensive tackles didn’t have any in the sample
reach the sample average for sacks or tackles for loss, only two eclipsing it in tackles.
The transition from college to the NFL being as directly correlated as it is
originates from job responsibilities staying uniform from college to the NFL. Scouts Inc.,
the preeminent recruiting service used for years by ESPN to supply high school football
rankings that were widely accepted as fact by the general public, states the top two duties
of defensive tackles that they look for when judging who will make a good player at the
position are toughness and control of the line of scrimmage (Scouts Inc., 2009), which
most often means beating the offensive lineman across from the prospect in question to
gain penetration towards the ballcarrier or quarterback. Dennis Dillon (2004) describes it
in the NFL as “lots of exertion, body trauma, and grunt work…plug the gaps on either
side of the center.” (Dillon, 2004, para. 5). With the duties being largely the same and
schematic changes from college to the NFL, unlike other positions, being relatively
minimal, seeing defensive tackles statistics from college translate to the NFL better than
any other position in this study is not a surprise, since the successes at the position have
been doing their job at a high level during college and require little adjustment in the
professional game. Other positions, even the offensive linemen discussed above that have
to stop the defensive tackles, have to learn new offenses and responsibilities. While there
are subtleties in defensive tackle play, it, unlike any other position, stays the same from
!
75!
college to the NFL, especially with teams drafting defensive tackles at the size needed to
fit their base defense.
Defensive Ends. In a virtual tie with quarterbacks and defensive tackles in our
sample for shortest career lifespan is the defensive end position, the 26 in our sample
lasting 6.84 years in the NFL. The conclusion one can draw from this fact is that
defensive ends, along with defensive tackles, are being hit by the largest athletes on the
field, offensive linemen, on every snap, leading to more injuries, quicker body
breakdown, and shorter careers. Quarterbacks, similarly, are standing relatively still
behind their offensive line and had a median time of 2.7 seconds to get rid of the ball in
the 2018 season (Passing, 2018) before being hit by an oncoming rusher that averages
over 300 pounds in weight. The target of men that size, it’s no surprise to see
quarterbacks in this conversation with defensive ends and defensive tackles. Still, the
6.84 years defensive ends taken in the first round from 2007-2012 average is more than
double the career expectancy of the normal NFL player, a number that dipped from 4.99
years to 2.66 in the middle of the 2010s (Arthur, 2016). Only two players in the sample at
the time of data collection had a career last double figure seasons, those being Brian
Orakpo, whose prolific production placed him third in the positional sample in NFL
tackles, sixth in sacks, and fourth in passes defensed, and Chris Long, who led the sample
in NFL games played. These numbers show why they made it 10-plus years, Orakpo
retiring after the 2018 season, Long soldiering on. Three others concluded the 2018
season with nine years of service, including the sample leader in tackles and passes
defensed, Jason Pierre-Paul. Drafting teams struggled to identify worthy first-round picks
at this position early in the sample, only Long and Tyson Jackson of the first nine ends
!
76!
chosen remaining in the NFL after six seasons. Gaines Adams, the second-highest draftee
in the sample, and Vernon Gholston, the third-highest who logged lows in NFL games
played, tackles, sacks, forced fumbles, and passes defensed, lasted just three seasons
apiece and were amongst the aforementioned first nine. Since then, only three of the
ensuing 17 first-round picks have not made it to the sample average of career length,
teams showing a penchant for better decision-making at the position once following the
ill-advised drafting of Aaron Maybin by the Buffalo Bills. Good arguments can be made
that Long, Pierre-Paul, J.J. Watt, the sample leader in NFL sacks and fumble recoveries,
Ryan Kerrigan, the sample leader in NFL forced fumbles and interceptions, and Cameron
Jordan have been the most impactful NFL ends of this six-year period.
Pierre-Paul accomplished this despite playing just 13 games of NCAA Division I
football, transferring to South Florida after two years at a junior college. Because of this,
Pierre-Paul ranks last in collegiate games played, tackles, and sacks amongst the sample.
Conversely, Kerrigan led his 25 counterparts in collegiate tackles, sacks, tackles for loss,
and forced fumbles. Between the two were collegiate performances that impressed, like
Lawrence Jackson of USC, who finished second in sacks and tackles for a loss before
going on to have a pedestrian five-year pro career, and those that lacked, Adams
averaging just two tackles per game and generating little in the way of impactful plays
over four years with Clemson before just three seasons in the NFL with Tampa Bay.
At the Combine, two that have nearly identical statistical resumes in the NFL put
up the low and high numbers in the 40-yard dash, Bruce Irvin running a 4.41 while
Cameron Heyward ran a 4.95, the two drafted seven picks apart in 2011. Adams was .05
seconds from matching Irvin, his 4.46 contributing heavily to Tampa Bay selecting him
!
77!
as highly as they did in 2007. Orakpo’s athleticism was displayed in the vertical jump,
leaping nearly 40 inches, 2.5 inches higher than any other defensive end in the sample,
while Jackson and Derrick Harvey mustered only 28.5 inches at the bottom of the
defensive end list. Gholston, whose Combine numbers were well-rounded in the three
areas measured in this study, benched the desired weight 37 times, tops amongst
prospects, while Jarvis Moss, who did not stand out in any area aside from height (6’6”
the tallest of any draftee) in his drills or measurements, lifted the weight only 16 times.
126 correlations were examined between the Combine as well as collegiate game
performance and their predictive validity towards NFL success. Pearson’s correlation
coefficient was the equation used to study the relationship, or lack thereof, between the
Combine and NFL performance as well as collegiate game play and NFL performance.
Of the 126 possible correlations, a meager five came back significant, three signifying a
positive relationship and two signifying a negative relationship.
Most notably, the height and weight measures suggested a certain body frame of
player is fit to be an NFL defensive end. Height, which negatively correlated to NFL
games and years played, was dwarfed by weight, which positively correlated with NFL
games played, leading to the conclusion that a tall, comparatively lanky defensive end
would not last a significant amount of time in the NFL. Already being outweighed by an
average of 36 pounds by offensive linemen (Blender, 2019), defensive ends that have a
higher center of gravity and, while being presumably more athletic than a shorter and
heavier defensive end, weigh less, run the risk of being severely outmuscled by the
heavier linemen across from them. The 2015 weights of select NFL defensive lines saw
Dallas have a combined weight of 243 pounds at their end positions, finishing 2015 in the
!
78!
bottom 10 in the NFL in rushing yards allowed, also allowing the fifth-most rushing
touchdowns and tying for the seventh-fewest sacks in the 32-team league. Also standing
out as being exceptionally slight at the end position was Miami, allowing the fifth-most
rushing yards that season while tying Dallas in quarterback sacks with 31 (“Ranking
NFL”, 2015). While these examples are neither complete nor exhaustive, it gives a real-
world snapshot into how lacking a stout figure at defensive end can hurt team success.
The sample showed the same, as players in the 26 defensive end group that
weighed 290 pounds or more all played in the NFL at least eight years, more than a year
above the sample’s career average. Edge rusher Elvis Dumervil, 6’0” in height, four
inches shorter than the average defensive end in the sample, had over 60 sacks in his six
first seasons with the Denver Broncos, and believes leverage is the reason height can
work against a player coming off the edge of the defensive line against an offensive
lineman:
"The bigger the lineman, the better. I don't like going against the shorter guys,
because I lose my leverage advantage. When I see the opponent's depth chart and
the offensive tackle is 6'7", I get way more excited.!I know at the end of the day,
he prepares for those defenders who are 6'4" or 6'5". He's got to get out of his
technique against me. He's accustomed to blocking taller players. When I get
lower, the blocker gets out of his comfort zone. That's when they have issues."
(Sobleski, 2016, para. 44).!
The scientific concept of leverage, a higher center of gravity vs. a heavier object,
and power, which is equal to force multiplied by velocity, help explain why a defensive
end with a larger mass given a running start would prefer to be heavy rather than tall.
!
79!
Linebackers. Aligned between the defensive line, tasked with stopping the run
and rushing the passer around the line of scrimmage, and defensive backs, asked to stop
the pass and guard eligible receivers, linebackers are required to do both, meaning the
multi-faceted position requires a hybrid athlete. Because of this, perhaps more than any
other position, the Combine demands strong performances from those that will play the
middle level of the defense. That is evidenced in Vincent, Blissmer, and Hatfield’s (2018)
study of the Combine from 2005-2010 in which linebackers accounted for the most
participants in the four drills the study measured, including 40-yard dash. In the study,
the 40-yard dash was significantly correlated with three of the four on-field statistical
categories measured. The concluding statement on defensive player success in the NFL
from Vincent et al.’s (2018) work is that the benefit of the current testing battery for
defensive player success was modest, but no player in this work’s sample was willing to
omit any of the measured drills aside from Jon Beason, who decided to forego the vertical
jump, and Donta Hightower, who declined the bench press. Linebackers, however, were
not alone in participating in every drill they could in this work’s sample, with all but 10
of the drills skipped coming from the offensive “skill positions” of wide receiver, running
back, and quarterback, which lends itself to differentiation between this study and
Vincent et al.’s (2018), in addition to the fact that the only like-drill that was measured in
the studies was the 40-yard dash.
The fastest in the 40-yard dash is the man that has gone on to the most dominating
NFL career, outside linebacker Von Miller, who led the sample in sacks and forced
fumbles. Miller’s specialty is rushing the passer, and he was named the top player in that
area across the NFL ahead of the 2018 season (Mcginest, 2018), speed to get around the
!
80!
outside edge against offensive tackles being integral in that dominance. Slowest in the
run at the Combine was Larry English, who led prospects in the sample in collegiate
sacks, but finished last in tackles once in the NFL of those measured. Leaping higher than
any other linebacker at the Combine was Keith Rivers, jumping 42 inches, though his
sacks and forced fumbles would prove to be last amongst the sample once he got to the
NFL, collecting just three and one, respectively. On the bench press, Nick Perry who
weighed in as the heaviest linebacker at 271lbs., reached 35 reps, helping to counteract a
career in which he had just under three tackles per game in college, Green Bay selecting
him as the last linebacker in the sample in 2012. Interestingly, Clay Matthews and Luke
Kuechly, two of the five most accomplished linebackers in the sample, found starkly
different success in college, Matthews finishing last in tackles for loss and tackles despite
playing the second most collegiate games in the sample, while Kuechly led the sample in
collegiate tackles, fumbles recovered, interceptions, and passes defensed. Kuechly has
been revered as the best linebacker in the league’s best core (Davis, 2017), while
Matthews had the longest active career at sample collection, trailing only Miller for NFL
sacks amongst the 19 researched.
126 correlations were examined between the Combine as well as collegiate game
performance and their predictive validity towards NFL success. Pearson’s correlation
coefficient was the equation used to study the relationship, or lack thereof, between the
Combine and NFL performance as well as collegiate game play and NFL performance.
Of the 126 possible correlations, only two returned a positive correlation, while 17, 14 of
which were from collegiate in-game performance, returned a negative correlation. The
only other negative correlations, weight at the combine forming a negative relationship
!
81!
with tackles, interceptions, and passes defensed, shows the need for a linebacker to be
agile in pass defense and to chase down ballcarriers, as heavier linebackers generally run
slower than smaller linebackers.
The 14 negative correlations saw seven of the nine collegiate in-game statistics
involved in negative relationships with six different NFL statistical categories measured
in the sample. With this wide-sweeping data covering the large majority of both
collegiate in-game performance and NFL success, the simple generalization that
exceptional collegiate statistics do not lead to strong NFL statistics can be stated. Those
that have had success in the sample can be split into two categories – pass rush
specialists, like Clay Matthews and Von Miller, and middle linebackers who amass huge
tackle numbers, like Luke Kuechly, Patrick Willis, and Lawrence Timmons. While some
benefitted from blazing speed, such as Miller and Willis, all five mentioned benefitted
from being lighter than the average linebacker in the study, which was 250 pounds,
Timmons the lightest at 234. One National Football Conference, one of the two
conference’s in the NFL, general manager addressed the shift away from heavier
linebackers that some teams have adopted: “Teams have gotten away from those big
‘backers because there's more emphasis on speed and blitzing and coverage and not
taking on blockers." (Sando, 2007, para. 7). While not all the linebackers listed above
will pile up sacks because of their role, and not all possess top-level speed, being able to
shift and not take on blockers is something that lends itself well to a man of less mass, as
is seen in the NFL statistics in the sample. Additionally, “size is only a plus if you are
athletic.” (Sando, 2007, para. 7). Indeed, if 40-yard dash has an interrelationship with
vertical jump (Mcgee & Burkett, 2003) that forms a cogent definition of athleticism, then
!
82!
even if a player does not run a top time in the 40-yard dash, they can help even out the
predictor of athleticism with a strong vertical leap. While Matthews, Timmons, and
Kuechly did not run standout times, they outperformed the average in vertical leap,
atoning for their relative sluggishness in the run. The need for more agile and athletic
linebackers is only to grow as the NFL continues to go towards more pass-heavy
offenses, with larger, less shifty, downhill runners at the linebacker outpaced by running
backs, slot receivers, but of most concern, more athletic tight ends that they could
previously cover effectively (Sando, 2007). The only thing, so it seems in the sample, that
can stop these smaller, athletic, shifty linebackers, is injury, as Sean Weatherspoon, Keith
Rivers, and Nick Perry put forth three of the four best vertical leaps amongst the 19
linebackers, and also had strong showings in the 40-yard dash, but averaged playing just
11 games of the 16 per year that are played in the NFL. The other top vertical leaps were
the ultra-productive Willis, Kuechly, and Miller. This data shows NFL franchises that
seeking out lighter linebackers whom show the speed, vertical jump, or a balanced
combination of both, and also display a level of durability to avoid injury, will lead to a
more successful linebacker than one that is simply productive in college.
Cornerbacks. Amongst the 20 cornerbacks in this study’s sample, three of them
were from non-Power 5 schools, tied with offensive linemen, which there were 16 more
of in the positional sample, for most amongst positions. Leodis McKelvin, Dominique
Rodgers-Cromartie, and Kyle Wilson made up the three, though of those, only one was
still active at the time of data collection, while they led two different NFL statistical
categories in the sample but finished last in four others. While the odds are stacked
against these prospects in terms of exposure, level of competition, and resources (Eckard,
!
83!
2018; Maisel, 2014), Abbasian, Sieben, and Gastauer (2016) find that players that attend
a non-Power 5 school face no disadvantage or bias in being drafted. In his work revolving
around high school football players and their “star” rating in recruiting, the authors
explored if players rated with different “stars” by independent recruiting services had a
better or worse likelihood of being drafted depending on which conference the school
they attended resided within. Abbasian et al. (2016) discovered that not only does the
particular school and conference not make a difference in being drafted, but it also
showed no difference in how early or late a player was drafted. While the sheer number
of draft picks from Power 5 schools was significantly higher, in some cases as many as
17 times the amount for a particular “star” rating, the study showed the non-Power 5
“starred” players in the draft from 2002-2013 did roughly as well, or better, in terms of
being drafted as well as what position they were taken. While that study shows the lack
of difference in conferences and schools for the aforementioned groups of prospects, this
study shows both the success, and the failure, the non-Power 5 group can have once they
arrive in the NFL, no position in this study speaking to this point quite like cornerback.
Whatever school the sample attended, the majority of those studied still have the
chance to improve their statistics, as cornerback is the position that averaged the longest
NFL lifespan in this study at nearly nine years. 14 of the cornerbacks’ careers were still
active, only two of the 20 total finishing their careers before they reached eight seasons.
Predicting their success, and longevity, off of college performance proved more
difficult than most positions, as the man to lead the sample in three separate collegiate
categories was one of the two men to last less than eight years in the NFL, Antoine
Cason. Aqib Talib, Devin McCourty, Leon Hall, and Stephon Gilmore led the other
!
84!
categories, while numbers say Morris Claiborne and Dre Kirkpatrick amassed the
statistically least impressive collegiate careers.
At the Combine, Dominique Rodgers-Cromartie and LSU’s Patrick Peterson ran
the fastest 40-yard dashes while also grabbing two of the top three heights in the vertical
jump. Mcgee and Burkett (2003) showed a strong correlation between the sprint drills
and the vertical jump at the Combine as interrelated measurements of athleticism, so to
see these numbers lead to Rodgers-Cromartie and Peterson having NFL success validates
their research, as does their conclusion that cornerback, running back, and wide receiver
are the three positions that Combine numbers tell the most about. Contrary to the vertical
and sprint drills, cornerback is not a position known for upper-body strength, and Kyle
Wilson, the player with the shortest NFL career, led the sample with 25 bench press
repetitions, though Vontae Davis, whose NFL numbers are rated above average in the
sample, also had 25.
126 correlations were examined between the Combine as well as collegiate game
performance and their predictive validity towards NFL success. Pearson’s correlation
coefficient was the equation used to study the relationship, or lack thereof, between the
Combine and NFL performance as well as collegiate game play and NFL performance.
Of the 126 possible correlations, 10 returned a positive relationship, while six returned a
negative relationship.
Most notably, of the 10 positive correlations, the similarities between offensive
linemen and cornerbacks continued, as corners joined offensive linemen as the two
positions to have definitive results showing Combine drills as a predictor of NFL success.
The aforementioned offensive linemen saw 40-yard dash predict three NFL categories
!
85!
positively, vertical jump positively correlating two, while cornerbacks raised those
numbers, positively correlating 40-yard dash to NFL games played and vertical jump to
six different NFL statistical categories, for seven total positive correlations.
Historically tasked with defending a large amount of deep passes (Mays, 2017),
cornerbacks, who line up opposite wide receivers, have been given the job of winning
one-on-one battles down the field on passes being dropped in from extreme vertical
angles that require impressive leapers to break up the passes. With those responsibilities
on this position, vertical jump correlating strongly to NFL success logically follows.
Corners, though, to be able to make the play on the deep ball, which can be game-
changing if successful, must be able to keep up with the receiver across from them in
order to be in a position to break up the pass. In order to do this, they need a strong 40-
yard dash time, as receivers and cornerbacks are the positions running in a straight line
for an extended distance most often. To that point, the 40-yard dash correlating positively
to NFL games played makes sense as well, as speed is difficult to gain if it is not already
possessed by a player, while technique and fundamentals are more easily taught. Those
with a strong 40-yard dash therefore are more likely to be given additional opportunities
in the NFL because their quality is more rare to find, and these players are also more
likely to succeed according to this study. Wide receiver versus cornerback evolution will
be worth monitoring, as wide receivers, discussed above, are having success when bigger
and stronger, and NFL teams are throwing the ball downfield less, opting for the more
high percentage short passing game (Clark, 2018). To this point, cornerbacks that test
well at the traditional responsibilities of their position are showing those areas still count,
however, with the most successful receivers in the NFL being the largest, as shown
!
86!
earlier in the sample, and the largest bodies on the football field residing closest to the
line of scrimmage, NFL teams using more short passing routes may lead to leaping and
straight-line speed becoming diminished qualities to possess at this position, while
strength and agility could be more in-demand. This, though, is not the case in this work,
as the traditional cornerback still has an important and prevalent place on the field.
Safeties. The last line of defense in football, safeties play furthest from the line of
scrimmage and are cornerback’s partners in the defensive secondary. Despite their
importance on the field, not a single safety was amongst the 50 highest-paid players
during the 2017 season (Gunter, 2017), the only other positions measured in this study
having that dubious distinction being running back and tight end. Perhaps even more
damning to safeties’ worth were the findings of Kutz (2017) when examining the highest
paid player at every NFL position, with the highest paid running back, Le’Veon Bell, and
the highest paid tight end, Jason Witten, ranking above the league’s highest-paid safeties
at the strong and free positions, Barry Church and Devin McCourty, respectively. The
only positions garnering lower numbers for their top-earners were Kyle Juszczyk at
fullback, a specialty position that is no longer used in the NFL by many teams because of
teams’ propensity to pass, and Jared Veldheef at right guard, making slightly below
McCourty but well clear of Juszczyk. With the average first-round pick getting a $16
million deal in 2018 (Belzer, 2018) and ensuing rounds getting at-most $5.3 million
contracts (Gaines & Yukari, 2017), the financial importance of being a first-round
selection speaks for itself. The league’s teams are also speaking for themselves with their
selections by position, as only 10 safeties were drafted from 2007-2012 in the first round,
leaving many of the best at the position with contracts a fraction the size of others that
!
87!
may have ranked lower at their respective position, but were playing a more desirable
role in the eyes of drafting teams. The only position with fewer draftees, as discussed
earlier, was tight ends, with four taken during this study’s sample years. With the
demands of the tight end position in flux and running backs used less in the modern
passing league that the NFL currently employs, their wage scale and diminished value
amongst positions seems to follow, though safeties seem undervalued considering they
are under fire now more than ever with the prevalence of passing. Sports analytics giant
FiveThirtyEight expands:
“Teams are certainly passing more often than they used to. Leaguewide passing
attempts per game have risen from 32.3 in 2008 to 34.2 last year, and the increase
in volume has not been accompanied by a decrease in efficiency. Leaguewide
yards per attempt have increased slightly from 6.9 to 7.0, and more touchdowns
are being scored by passing relative to running than at any time in league history.
Completion percentage is up from 61.0 percent to 62.1 percent, and the
interception rate has fallen from 2.8 percent to 2.5 percent.” (Hersmeyer, 2018,
para. 4)
Further, one of the players in the safety sample was drafted as a cornerback,
Malcolm Jenkins, before being moved to safety after he was drafted. Still, Jenkins is
included in the safety position because he only played one season at cornerback in the
NFL, and to normalize stats, as they skewed heavily towards that of a safety, with more
tackles and sacks, as opposed to a cornerback, who would have more passes defensed.
Devin McCourty, discussed in the cornerback sample, spent more time than Jenkins at
!
88!
the cornerback position in the NFL before moving to safety, so his stats were kept with
cornerbacks.
Of those remaining in the 10-safety group, Laron Landry ran the fastest 40-yard
dash, a 4.35, while Eric Berry, just five hundredths of a second behind Landry, lept 43
inches, 5.5 higher than Landry’s vertical jump that was good for second in the sample.
Landry and Berry were thus the two highest-drafted safeties, though both had off-field
issues derail their careers, Landry suspended indefinitely after eight NFL seasons for a
violation of the NFL’s performance-enhancing drug policy (Bieler, 2015), Berry
sidelined by, amongst other less serious injuries, a bout with cancer, during his eight-year
career that he hopes to continue (Teicher, 2018). The safeties drafted most recently in the
sample, Harrison Smith and Mark Barron, ran 4.54 40-yard dashes, slowest in the group,
while also logging two of the four shortest vertical leaps, but have had careers on-par
with the rest of the sample. The only player whose career lasted fewer than eight seasons
and is no longer active amongst the 10 studied was Kenny Phillips, whose vertical leap
and 40-yard dash were lacking, while college stats were middling. Phillips lasted only six
NFL seasons and was last in five of the seven statistical categories throughout the sample
when in the NFL.
Collegiately, Michael Griffin paced the group in three of the eight statistical
categories before posting average Combine numbers, while Reggie Nelson finished last
in six of the eight, his numbers at the Combine resembling Griffin’s closely. Barron led
the sample in collegiate games played, though his production did not follow, generating
substandard collegiate stats comparatively. Landry led the sample in sacks and passes
defensed, Brandon Merriweather in tackles for a loss, and Berry in interceptions.
!
89!
126 correlations were examined between the Combine as well as collegiate game
performance and their predictive validity towards NFL success. Pearson’s correlation
coefficient was the equation used to study the relationship, or lack thereof, between the
Combine and NFL performance as well as collegiate game play and NFL performance.
Of the 126 possible correlations, four returned positive correlations, while seven returned
negative correlations.
Most notably, the negative correlations rejected both height and weight at the
Combine predicting NFL success, negatively correlating with both NFL games and years
played, along with NFL passes defensed. This was the only position group in which both
height and weight were negatively correlated with NFL success, wide receivers being
positively correlated with both. Conversely with the other defensive secondary position,
cornerback, which is striving to be larger to match the growing size of receivers (Birkett,
2017), the work done in this study shows safeties assuming the traditional role of
cornerbacks – to be able to cover lots of ground and keep up with receivers on long
passing plays, lending the position to a smaller player that can more easily change
direction and adjust his mass who will need to play further from the line of scrimmage, as
opposed to some safeties that have historically played more towards the defensive line
and linebackers to act as a quasi-extra linebacker. The ability to do a wide variety of
things for the modern safety has been a key, and while straight-line speed did not
correlate to NFL success with safeties, their average vertical jump was the highest on the
defensive side of the ball, only wide receivers offensively topping the 35.9 average
inches leapt, showing that athleticism is high across the sample at the position. This is
important because of jump balls downfield and the need to compete with the statistically
!
90!
top athletes in the sample, wide receivers. The emphasis seems to be shifting from
safeties being the more physical, hard-hitting of the two defensive secondary positions to
cornerbacks being the more physical of the two, which follows the logic discussed in the
wide receiver and cornerback portions of this results section stating that NFL teams are
throwing more short passes, meaning those closer to the line of scrimmage need to be
able to fight through bigger bodies to make plays. Should this trend continue, the results
in this study show that safeties would turn back into the players they were designed to be
at the outset of football’s modern defensive schemes – the last line of defense that needs
to be able to jump with receivers and stop the big passing play. Andy Benoit (2014) of
Sports Illustrated discusses the multitude of responsibilities safeties may have:
“Offensive coordinators have been widening the field and featuring athletic tight
ends to create favorable mismatches against safeties. Often run at no-huddle
tempo, spread sets typically require man-to-man coverage, which means defenses
are demanding more out of their safeties. Cornerbacks and pass rushers remain
essential. If your corners can’t cover, you’re hamstrung. If your pass rushers can’t
generate pressure, you’re playing uphill. But if you don’t have quality safeties,
you’re ultimately at a creative disadvantage.” (Benoit, 2014, para. 4)
Defense is clearly a team effort, and one slight adjustment to Benoit’s reasoning
may be vital in the coming years, with cornerbacks still needing to cover, but also
needing to be physical and attack the short passing game so safeties can patrol the back
half of the defensive secondary. Safeties have often needed to be the do-it-all player on
defense, but this study reveals that, while having a hard-hitting, fast, intelligent safety is
ideal, if you can pick only the most important quality, finding the smaller, shiftier safety
!
91!
and putting your larger, more physical players at the cornerback position sets a defense
up for success. The size matchup will continue to intensify between wide receivers and
cornerbacks, while safeties show the trend of turning into more of a pass support role that
NFL defenses need in the modern day.
Power 5, Big East and Notre Dame v Non-Power 5. The goal of this portion of the
author’s work is to determine if attending a school inside the Power 5, plus the Big East
Conference, and independent Notre Dame, is predictive of NFL success, or if attending a
non-Power 5 school gives a prospect as much of a chance to have a prolific NFL career.
The process for determining which schools, and therefore conferences, both inside and
outside of the Power 5, were the strongest at the time of the sample was important to
ensure schools were being ranked appropriately. To do this, the authors used the Sagarin
Rankings, which represent the average schedule difficulty faced by each team in the
games that they played to a given point in the season, the schedule difficulty of a game
taking into account the rating of the opponent and the location of the game (Sagarin,
2019). The rankings are a combination of two computer-based point systems that are used
to generate a final rating, and also log college football teams’ record against top-10 and
top-30 teams in the rankings. For nearly 50 years, these rankings have been a part of the
mainstream consciousness and one of the most trusted ways to rank college football
teams (Feng, 2019).
After the Sagarin Rankings were selected and reviewed for legitimacy, the question
arose of what timeframe to measure. In order to ensure every prospect in the 2007-2012
sample was accounted for, the period of Sagarin Rankings collected was 2004-2011, the
timeframe adjusted to cover prospects that entered the 2007 draft and would have played
!
92!
collegiately during the 2004-2006 seasons. In collecting this data, the final rankings from
each season were the ones acquired, logged, and averaged to determine where each
school rated in relation to 238 others that the rankings covered over the eight-year period.
This method of data collection and determination of top teams was successful in
when comparing it to a few other ways of doing so. The AP Poll, for decades the most
trusted arbiter of college football’s national champion (Tracy, 2016), had 190 of the 200
top-10 slots over a two-decade period occupied by Power 5 teams, equaling 95 percent of
those spots available (Eckard, 2018). Over the eight years of collection for the Sagarin
Rankings in this study, nine of the 10 teams in the top-10 came from Power 5
conferences, 90 percent, the exception being Boise State, who ranked ninth thanks to
their status as one of the most improbable Cinderella stories in NCAA history (Fornelli,
2017). Further, 23 of the top 25 programs in the averaged Sagarin Ratings were from
Power 5’s, 92 percent, inching closer to Eckard’s (2018) AP Poll number. Further
reinforcement comes from the national championship game, which during the Sagarin
Rankings collection timeframe pitted the nation’s top two teams in the BCS rankings
against each other to determine the best team in the country. Over the eight years of
collection, nine programs played in the game - USC (2), Texas (2), Oklahoma (2), Florida
(2), Ohio State (2), LSU (2), Alabama (2), Auburn (1), Oregon (1) – seven different ones
winning it with only Alabama achieving multiple titles. The position of the programs
appearing in the national championship game during that time in the Sagarin Rankings
average of the 65 schools that data was collected on was first, second, third, fifth, sixth,
seventh, eighth, 11
th
, and 13
th
. Oregon and Alabama were 8
th
and 13
th
, respectively, only
halted from being higher in the rankings because of the first few years of the collection
!
93!
period, both going through a coaching transition that would lead to the programs
ascending in the latter half of the sample, Alabama on their way to a dynasty that has
garnered five national championships in the last 10 years. Additionally, the schools
outside Notre Dame, the Power 5, and the Big East had an average program rating over
the eight years of 94.95 and an average rating amongst the 65 schools to have a player
drafted of 52.16, only brought to that peak by Boise State and TCU who ranked in the
top-20, the non-Power 5 teams filling nine of the bottom 10 positions in the 65-team
rankings.
The program with the best score over the eight years measured was the PAC-12’s
University of Southern California, averaging 8.125. USC also accounted for half of the
20 prospects drafted during the 2007-2012 sample period, 10 ranking second amongst the
65 schools to have a player drafted during the six-year sample, Alabama garnering one
more first-round draftee. Despite ranking fifth across conferences with only 20 prospects
drafted during the sample period, the PAC-12 did place three schools in the top quarter of
the 65 schools average scores from 2004-2011.
Placing second amongst the 65 programs was the BIG-12’s University of Oklahoma,
averaging 8.375. The Sooners tied for seventh amongst the measured institutions in
prospects drafted during the sample with five, the BIG-12 tying the Southeastern
Conference for the most individual programs to have a player drafted with 11. The BIG-
12 finished behind only the SEC for most prospects drafted, sending 34 first-rounders to
the NFL over the six years of the sample, the SEC accounting for more than a quarter of
the 191 prospects in the sample, claiming 51 first-round draftees.
!
94!
The top SEC school in the program rankings was LSU, finishing third overall
amongst the 65 schools while averaging a score of 9.125. The Tigers finished second in
the conference in first-round picks, the aforementioned Alabama leading the conference
and the entire sample with 11, while the SEC placed five programs in the top 13 in
average score over the eight years of Sagarin Ratings collection. The proliferation of SEC
schools towards the top helps qualify the conference Sagarin Rankings, also kept by
MIT-graduate and founder Jeff Sagarin in addition to his individual program rankings.
Since the conclusion of the 2005 rankings, the SEC has finished in the top spot in the
final poll eight of a possible 13 times. No other conference in the 15 years since this work
began to average the Sagarin Rankings in 2004 has had any string of dominance even
approaching the SEC’s, the four other Power 5 conferences (ACC, BIG 12, BIG TEN,
PAC-12) logging either one or two years at the top of the rankings to make up the seven
years the SEC did not take the top spot.
In-depth calculations are not needed to determine whether attending a Power 5 school
gives a prospect a higher chance of being selected in the first-round, as the most basic
numbers act as overwhelming evidence on their own. Of the 191 players drafted in the
first round from 2007-2012, 176 of them attended either independent Notre Dame, long
considered one of the best programs in the country, an SEC, ACC, BIG TEN, BIG-12,
PAC-12 or Big East program. The split amongst non-Notre Dame, Power 5, or Big East
attendees broke down into three from the Mountain West Conference (MWC), three from
the Mid-American Conference (MAC), three from the Western Athletic Conference
(WAC), two from Conference-USA (C-USA), one from the Big Sky Conference, one
from the Sun Belt Conference, one from the Ohio Valley Conference (OVC), and one
!
95!
from the Colonial Athletic Conference (CAA). The sheer number, over 90 percent in the
sample, of Power 5, Big East, and Notre Dame players selected in comparison to those
outside of this group, seems to introduce bias and an unfair advantage to those in the
majority, as suggested by Ivan Maisel (2014) regarding the vote of autonomy that passed
for Power 5 schools referenced towards the end of the Methods section of this work.
Abbasian et al. (2016), though, found that players that attend a non-Power 5 school face
no disadvantage or bias in being drafted. Abbasian et al. (2016) discovered that not only
does the particular school and conference not make a difference in being drafted, but it
also showed no difference in how early or late a player was drafted. While the sheer
number of draft picks from Power 5 schools was significantly higher, in some cases as
many as 17 times so, the study showed the non-Power 5 players in the draft from 2002-
2013 did roughly as well, or better, in terms of being drafted as well as what position they
were taken in the draft. The distinction in Abbasian et al.’s (2016) study is an important
one – non-Power 5 prospects are at a disadvantage in being drafted, but that is not the
case when factoring in proportionality of those that are drafted for each group.
Rather than dissect whether attending a Power 5 increases the chances of being
drafted in the first round, as in sheer number it clearly does, Research Question No. 2
focuses on if players that attend a Power 5 conference program have a better chance of
NFL success. In order to do this, two separate data sets were examined and correlated
with the average score of the university the sample attended. The first set measured
university attended against collegiate game performance to determine whether those that
attend non-Power 5 schools put up disproportionately larger statistics because of facing
inferior competition. This would mean that, in future studies, weighting data would be
!
96!
important in trying to correlate collegiate in-game performance to NFL success, leveling
Power 5 stats to non-Power 5 stats to have them on an even plane. Robbins (2010) has
tried to explore this previously with Combine data, using ratio-scaled data versus raw
data, and while his study had interesting findings, that aspect of it failed. The second data
set measured university attended against NFL statistics, both data sets broken down by
position, to see if the university a prospect spent their collegiate career gave them a better
chance of NFL success. Pearson’s correlation coefficient was the equation used to study
the relationship, or lack thereof, between university attended and collegiate in-game stats
as well as NFL statistics. In data set one, 52 correlations were calculated, while data set
two harbored 68 correlations.
In the first data set, only three correlations of the 52 calculated were significant,
with no consistent patterns emerging across the 10 positions that the sample was broken
into. As was the theme for this work, quarterback, running back, and wide receiver, the
most studied positions in studies similar to this one, showed little return, with only wide
receivers having a negative correlation between university attended and collegiate
receiving yards. This means that the better school the receiver attended, the more difficult
it was for him to amass that statistic. That though, was the only correlation of those three
positions in data set one addressing Research Question No. 2, joining linebackers, those
attending schools with better Sagarin scores logging fewer tackles for loss, and defensive
ends, who forced fewer fumbles the higher their school rated in the Sagarin Rankings
from 2004-2011. Most notable from the first data set is that there are no correlations
signifying that statistics are more difficult to achieve at non-Power 5 schools, though the
three significant correlations of 52, while they all show statistics are more difficult to
!
97!
procure at schools rated better in the Sagarin Rankings, equal just over five percent of the
correlations calculated, a very slim number.
In data set two, across 68 possible correlations, 11 were significant, four forming
a positive relationship, while seven formed a negative relationship. Similar to data set
one, zero correlations were significant in determining quarterback, running back, or wide
receiver success. The correlations, rather, came from the defensive side of the ball,
defensive ends seeing their NFL tackles for loss, sacks, and forced fumble statistics suffer
as they moved closer to the top of the Sagarin Rankings, cornerbacks finding it difficult
to amass NFL games and years played, passes defensed, and interceptions, while safeties
proved to be the only position that registered correlations between university attended
and NFL stats that indicated it was less difficult to amass stats in the NFL if the prospect
attended a program ranked higher in the Sagarin Rankings. It should be pointed out,
however, that the 10 safeties in the sample all attended Power 5 schools, many of them
highly successful ones, all ranking in the top-half of the 65-team rankings, an average
rank of 15.1 of the eight schools represented. This correlation, then, does not directly
indicate that safeties at non-Power 5 schools will be devoid of success in the NFL, only
that there is variance among Power 5 schools at the safety position.
!
98!
CHAPTER 5
DISCUSSION
Discussion of Findings
This study of NFL first-round draft picks from 2007-2012 set out to discover
which was more predictive of NFL success amongst the sample - the NFL Draft
Combine, an allegedly key determinant of selection in the NFL draft, or collegiate in-
game performance, the proving ground where prospects battle against each other for
program, and individual, success and accolades. This study was the first to measure only
first-round draftees as a population, doing so because of the enormous wage gap between
first-round picks and the rest of the draft. The tremendous financial burden attached to
being selected with one of the first 32 picks of each draft means the choices made by
NFL teams with their first-round picks is of heightened importance, and determining how
to select those players is of extreme significance. The study also explored if attending a
larger, more successful, highly recognized program in one of the Power 5 conferences,
the Big East included because of timeframe and Notre Dame included because of lineage
and past triumphs, is more indicative of NFL success. In order to delve into these topics,
the author used Pearson’s correlation coefficient to study the relationship, or lack thereof,
between the Combine and NFL performance, collegiate game play and NFL
performance, and school of choice and NFL performance. In Research Question No. 1,
the bench press, vertical jump, and 40-yard dash were the drills of measure at the
Combine, while varying statistical categories dependent on position were chosen to
quantify collegiate and NFL performance. In Research Question No. 2, in order to
determine which individual schools, and therefore conferences, were the strongest during
!
99!
the sample period, the authors turned to the widely-respected Sagarin Rankings which
represent the average schedule difficulty faced by each team in the games that it's played
to a given point in the season, the schedule difficulty of a game taking into account the
rating of the opponent and the location of the game (Sagarin, 2019). These rankings have
been a part of the mainstream consciousness for nearly 50 years and one of the most
trusted ways to rank college football teams (Feng, 2019), the final rating of each year
averaged from 2004-2011 for each of the 65 schools that had a player selected in the first-
round of the NFL draft from 2007-2012 in order to come up with a ranking for each team.
Research Question No. 1. 940 total correlations were equated with Pearson’s
correlation coefficient, broken down into quarterback, running back, wide receiver, tight
end, offensive line, defensive tackle, defensive end, linebacker, cornerback, and safety
position groups. 126 correlations were calculated for each of the five defensive positions,
while 84 were logged for quarterbacks, 66 for wide receivers, tight ends, and running
backs, and 28 for offensive linemen. Of the 940 total, 95 showed significance, 53 with a
positive relationship, 42 with a negative relationship. Quarterbacks showed a negative
relationship between vertical jump and NFL career length as well as statistical output in
key categories, suggesting that athleticism, based of Mcgee and Burkett’s (2003)
interrelational understanding of vertical leap, 40-yard dash, and athleticism, is not
important in having NFL success. Running backs rejected the findings of Lyons et al.
(2011) as well as Kuzmits and Adams (2008), as the backs in this sample did not show
40-yard dash times maintaining correlational significance to the position’s NFL success,
unlike the study done by the duo in 2008. The numbers did, however, support Robbins
(2010) and Park (2016) in the finding that 40-yard dash time most influenced NFL draft
!
100!
position amongst running backs, 15 of the 17 in the sample running faster than the 13-
year average of the drill across the entire draft (Doll, 2013). Wide receiver results suggest
that size and strength are especially important at the position, while the findings of
Mulholland and Jensen (2016), which supported speed as mildly predictive of NFL
success across their longitudinal study of receivers, were rejected. Tight ends, despite
having a representative sample of the population, over 11 percent included in this study,
needs further investigation, as only four were selected from 2007-2012, in-game
collegiate performance correlating negatively with NFL success for the smallest
positional group. Offensive linemen, despite having the least data to analyze, had the
highest positional percentage of correlations across the study, showing the importance of
linemen needing to be modern day athletes amongst the all-positive correlations gleaned
from the group. Defensive tackle had the most positive correlations of the 10 different
positions, showing a direct relationship with prolific collegiate performance and NFL
success stemming from the job responsibilities of the position being most similar from
college to the NFL across all positions measured. Amongst defensive positions, defensive
ends shed the most light on how body frame can affect NFL success, showing that a
heavier end was more likely to be on the path to a long and stable NFL career than one
that is a more tall, slim, and presumably athletic defensive end because of the scientific
concepts of power, leverage, and gravity center. The linebacker sample rejected the
premise that collegiate in-game performance leads to NFL success for the position, while
it showed a need for a lighter, more agile, athletic linebacker who can shift quickly but
also be durable and reliable. Cornerbacks, given positional responsibility, had the most
expected of results, with the Combine 40-yard dash and vertical jump correlating
!
101!
positively to NFL success, though this is a position that is going through a fundamental
shift with their counterpart in the secondary, that being safeties, whose NFL success
correlated negatively with height and weight. Those two positions, in an NFL that is
relying much more on the pass (Hersmeyer, 2018), are taking on different roles, teams
looking to add size at the cornerback position to help defend an increasing amount of
short passes (Clark, 2018), while successful safeties are growing lighter and smaller to be
able to cover much of the defensive secondary that is becoming theirs more and more
with every passing season, and every opponent’s passing attempt.
With the varying positional results, this study rejects all generalizations about
Combine drills being predictive of NFL success regardless of position. With the in-depth
positional look at all 10 groups inspected in this work, it should be overwhelmingly
apparent that each position on an NFL team requires different things and must be
evaluated independently of one another. As an example, this work supports the
athletically inclined drills studied by the author, those being 40-yard dash and vertical
jump, as predictive of NFL success for cornerbacks and offensive linemen, though the
work done by Lyons et al. (2011), Kuzmits and Adams (2008), and Mulholland and
Jensen (2014, 2016) that suggested varying skills positions, those being running backs,
wide receivers, and tight ends, will have success based off a strong 40-yard dash at the
Combine, are rejected. One aspect of Lyons et al.’s (2011) work not discussed to this
point is supported by this study, as the authors of that study did a post-hoc analysis of the
bench press, finding that it does not predict NFL success. The author agrees with this
conclusion, as bench press correlated just once throughout the calculation of the collected
secondary data, that a negative correlation to NFL success at the cornerback position.
!
102!
Overall, 19 of the 95 total correlations that showed significance in this study came from
the Combine drills, 17 emerged from the height and weight measurements, and 59 came
from collegiate in-game performance. Of the 95 correlations, 70 were from the defensive
side of the ball, nearly triple the amount of correlations sprouting from the offensive
positions, with only 25 found amongst quarterbacks, running backs, wide receivers, tight
ends, and offensive linemen, nine coming from the latter. These numbers help this study
support Lyons et al.’s (2011) conclusion that “samples”, in this case collegiate in-game
performance, are more predictive than “signs”, those being Combine drills. Height and
weight accounting for the other 17 correlations was an unexpected event, but can help
bolster, and extend, Teramoto, Cross, and Willick’s (2016) study, as well as Mcgee and
Burkett (2003) work, Teramoto claiming height was most predictive of NFL success for
wide receivers, Mcgee and Burkett (2003) finding height and weight were predictive of a
receiver being drafted. This study concurs with Teramoto et al. (2013), adding that
weight is right alongside height, if not exceeding it, in predictive validity at the receiver
position, while this work furthers the assertion from Mcgee and Burkett (2003) about
draft position and adds that NFL success at the receiver position is predicted by height
and weight as well. While this study did not scrutinize the relationship between Combine
drills and being drafted, a tertiary finding of the study supports that 40-yard dash can
predict a positive result in the NFL draft, with 15 of the 17 running backs in this study
that were selected in the first round from 2007-2012 running better than the average 40-
yard dash time at running back already in the NFL from 2000-2012 of 4.49 (Doll, 2013).
This reinforces the data put forth by Robbins (2010), Park (2016) and Mcgee and
!
103!
Burkett’s (2003) studies that focused more extensively on the predictive validity of draft
success coming from Combine performance.
Research Question No. 1’s ultimate goal was to determine which was a better
predictor of NFL success, and as mentioned above, collegiate in-game performance
ultimately proved to be the more predictive of the two. While that fact cannot be
disputed, just how predictive either actually is, at least measured with the drill and
statistical testing battery used in this study, is highly questionable. With 940 correlations
possible and only 95 correlating significantly, 10.1 percent in total, that including height
and weight correlations that were not thought to be a focal point of either data set focused
on in this study, the temptation is to suggest a drastic overhaul of scouting practices,
determinants of drafting decisions amongst first-round selections, and a new Combine
testing battery, as many that have done studies similar to this have (Kuzmits & Adams,
2008; Robbins, 2010; Vincent et al., 2018). This is especially the case considering the
most optimistic attitude of any study contained within this work was Vincent et al. (2018)
who claimed the Combine as a “modest” predictor of NFL success, this study only
beginning to approach that level of buoyancy about the Combine’s drill effectiveness.
But rather than suggest a new testing battery or dismantle the event as a sham, this study
does not overreact to the lack of predictive validity amongst the Combine drills, the three
most visible and recognized amongst those performed by prospects looked at in this
work, contrastingly encouraging fans, academicians, and media to view the Combine as a
media event created by the National Football League to keep its brand relevant in a time
which the organization may otherwise fade from the collective sports consciousness.
With the Super Bowl having concluded the league’s schedule three weeks earlier, Major
!
104!
League Baseball starting its spring training, the National Basketball Association and
National Hockey League pushing towards their postseasons, and college basketball
approaching the conclusion of its regular season and beginning of its trademark event,
“March Madness”, the NFL could otherwise be forgotten about. Instead, wisely placed at
the end of February in the lead up to late April’s NFL Draft, the televised workouts allow
a look at players that fans may know from their college careers, speculation to build
around the work done by prospects at the Combine and their ensuing Pro Days a few
weeks later, and mock drafts from experts to prognosticate the results of the late spring
event. Rather than a failure of predicting NFL success, the Combine is a booming success
of brand and media strategy by the most profitable sports league in the United States.
Monitoring if this strategy continues to be strong going forward with the emergence of
the Alliance of American Football (AAF) in 2019, which begins play right after the NFL
season concludes, and the Xtreme Football League (XFL), yet to release a schedule at the
time of this work but slated for a 2020 kickoff, will be intriguing, with a portion of the
Combine moved to network television for the first time in 2019 (Hofheimer, 2019).
While this study’s work on the Combine resembled Kuzmits and Adams (2008)
findings for predictive validity, both studies gaining significance on roughly 10 percent
of the possible correlations calculated, there were three or more significant positive
correlations between in-game collegiate performance and NFL success for four of the
five positions on the defensive side of the ball, defensive tackles posting 11, linebackers
not far behind with 8. In the secondary, there were seven positive correlations, therefore
this study can partially support Mcgee and Burkett’s (2003) work, though the findings
were not consistent enough to fully do so. Rather, safeties showed more conclusive
!
105!
measurements between height and weight to NFL success, cornerbacks revealing rather
definitive evidence of athletic drills predicting their NFL outcome. Considering the
drastic difference in the number of positively correlated defensive statistics predicting
NFL success to the miniscule amount of statistics doing so for the skill positions on the
offensive side of the ball, this study shows the traditional stats kept on defenders still
serve as a good measuring stick in the college game and can be trusted for predicting
NFL success. Contrarily, with just three positive correlations for the skill position players
from collegiate in-game performance, the traditional statistics examined in this study
cannot be trusted, and in evaluation of quarterbacks, running backs, wide receivers, and
tight ends, must be abandoned. Previous studies have not been nearly as transparent with
their statistical testing battery as the current work, and in so doing, this study hopes to
allow academicians and talent evaluators the understanding of what works and what does
not in judging predictive validity in statistics, and will hopefully lead others to be as
transparent with their statistics, many previously not doing so likely hurting the past
advancement of NFL statistic development. Now that some of that development has
happened in the modern day sabermetric statistic era that sports is currently in, plus the
discrediting of traditional statistics has occurred in this study amongst the oft-evaluted
skill position players, there is ample opportunity to study alternative, in-depth, and
exploratory statistics. Some options are put forth in the Limitations and Directions for
Future Research portion of this work.
!
106!
Table 5.
Research Question No. 1 Significant P-Values – Offensive Positions
Variable | NFL | NFL | NFL | NFL | NFL | NFL | NFL | NFL | NFL | NFL
Years Games Pass Rush Total Recep- Receiving Yards Receiving Pro
Played Played Yards Yards TD tions Yards Per Catch TD Bowls
40-Yard -.466** -.537** -.523**
Dash (OL) (OL) (OL)
Vertical -.717** -.642** -.621** -.545* .385*
Jump (QB) (QB) (QB) (QB) (OL)
.334*
(OL)
Bench
Height @ .674**
Combine (WR)
Weight @ .460* .463* .478*
Combine (WR) (WR) (WR)
College .398* .427** -.664**
Years (OL) (OL) (QB)
Played
College .434** -.960* -.968* -.963* -.995**
Games (OL) (TE) (TE) (TE) (TE)
Played .476**
(OL)
College .489*
Yards Per (QB)
Attempt
College .488*
Rush (RB)
Yards
College -.952*
Receptions (TE)
!
107!
Table 6.
Research Question No. 1 Significant P-Values – Defensive Positions
Variable | NFL | NFL | NFL | NFL | NFL | NFL | NFL | NFL | NFL
Years Games Tackles Tackles Sacks Fumbles Fumbles Intercep- Passes
Played Played For Loss Forced Recovered tions Defensed
40-Yard -.553* -.613** -.512**
Dash (CB) (DT) (DE)
Vertical .615** .501* .547* .473* .614** .632**
Jump (CB) (CB) (CB) (CB) (CB) (CB)
Bench -.450*
(CB)
Height @ -.413* -.423* -.748* -.644*
Combine (DE) (DE) (S) (S)
-.704* -.636*
(S) (S)
Weight @ -.785** .760** -.650** -.500* -.669** -.633**
Combine (S) (DE) (LB) (DT) (LB) (LB)
-.708*
(S)
College -.493* .636* -.550*
Years (LB) (S) (LB)
Played
College -.541* .648* .690* -.466*
Games (LB) (S) (S) (LB)
Played
College -.477* .441* -.525* .609**
Tackles (LB) (DT) (LB) (CB)
College .483* .520* .440* .548* .454*
Tackles (DT) (DT) (DT) (CB) (DT)
For Loss -.668*
(S)
College -.539* .476* .444* .458* -.585** .450*
Sacks (LB) (DT) (DT) (LB) (LB) (DT)
.477* -.495**
(LB) (LB)
.705*
(S)
College .571** .573** .443* -.469* .493* -.518* .439*
Fumbles (DT) (DT) (DT) (CB) (DE) (LB) (DT)
Forced -.510* -.544* -.503* -.531* -.466*
(CB) (CB) (LB) (CB) (CB)
College .447*
Fumbles (CB)
Recovered
!
108!
College -.566* -.474*
Intercep- (LB) (LB)
tions
College .443* .471* .547* .481* .654**
Passes (DT) (DT) (DT) (DT) (DT)
Defensed -.465*
(LB)
Research Question No. 2. 120 total correlations were calculated in two different
data sets. In the first data set, the goal was to find a relationship, or lack there of, between
amassing large collegiate stats and the program a prospect attended in college. This
meant that the Sagarin Rating average that was calculated from 2004-2011 by the author
was examined against the in-game collegiate statistics the prospect accrued to determine
if statistics amongst first-round picks got smaller as the prospects in question attended
better schools. That, though, was not the case. Once again broken into positional groups,
of the 52 possible correlations within the calculation of data set one, only three were
significant, and while they all indicated that statistics were more difficult to stockpile at
schools rated better by the Sagarin Ranking, the significance was isolated and did not
withstand scrutiny, showing that the top players are just as likely to build statistics at a
Power 5 school as opposed to a non-Power 5 school. In the larger picture, this shows that
it is not appropriate to normalize or manipulate data, but rather it should be taken at face
value because level of competition did not have an impact on the statistics amassed by
first-round picks. Robbins (2010) used ratio-scaled and allometric tactics to attempt to
normalize Combine data to produce a better predictor of draft success, and that element
of Robbins study failed. With the findings of collegiate statistics not being affected by
school, this study shows similar methods would not be necessary with collegiate in-game
performance data.
!
109!
The second data set cross-examined the same Sagarin Ranking calculated in the
first data set of Research Question No. 2 with NFL in-game statistics to determine if
attending a school with a better ranking, therefore facing better competition, would lead
to greater NFL success. This time, the results were mixed. Of the 11 correlations, seven,
at the defensive end and cornerback positions, showed NFL statistics actually suffered as
the Sagarin Ranking vaulted towards to the top of the 65-team sample. The other four
correlations were at the safety position, and showing that the better competition a
prospect consistently faced the more ready he was for the NFL, the better Sagarin
Ranking led to more NFL games played, tackles, forced fumbles, and fumbles recovered
for the players. It is worth noting that none of the safeties attended schools outside of the
Power 5, so while a relationship cannot be directly drawn from the sample to non-Power
5 prospects, it can be inferred that if the Sagarin Ranking continued to drop, the safety
would be less likely to succeed in the NFL. Quite simply, for defensive ends and
cornerbacks, the worse the school’s collegiate competition the better the player ended up
being in the NFL, and with safety population, the better program the player attended, the
better he turned out in the NFL. While the result is split, and seemingly leaves some room
for doubt, in reality it does the opposite because of the variance of positional results. This
casts more certainty on Abbasian et al.’s (2016) findings that non-Power 5 schools faced
no disadvantage or bias in being drafted, while the position in which they were taken
during their sample also was not affected by the conference their school resided within.
This study comes to the same conclusion, with the exception of the safety position, and
adds that NFL success can be found in a variety of different places throughout college
football. The Pearson correlation coefficient shows that through this work, and the
!
110!
individual players do as well, with Chris Johnson, one of only seven athletes in the
history of the NFL to rush for over 2,000 yards in a season, Joe Flacco, who has thrown
for the 22
nd
-most career yards in NFL history, and six-time Pro Bowler Joe Staley
headlining the 15 non-Power 5 players in the sample. The three correlations in data set
one indicate that players from all different walks of college football can assemble
noteworthy statistics, and while it may be tempting to forego positional analysis because
of that, data set two shows that, much like Research Question No. 1, the truly definitive
significance in this study lies on the defensive side of the ball, showing that positional
specification is needed in Power 5 v non-Power 5 studies as well. Within that positional
analysis, no weighting of statistics is needed because of the findings within data set one
discussed above, though skill position results once again speak loudly by providing no
significance at all, a need for alternative examination of quarterbacks, running backs,
wide receivers, and tight ends necessary to verify the findings in this study.
!
111!
Table 7.
Research Question No. 2 Significant P-Values – Power 5 v. Non-Power 5
Variable Non-Power 5------------------------University Attended----------------------------Power 5
College .445*
Receiving (WR)
Yards
College .630**
Tackles (LB)
For Loss
College .499**
Fumbles (DE)
Forced
NFL .554* -.675*
Games (CB) (S)
Played
NFL .464*
Years (CB)
Played
NFL -.836**
Tackles (S)
NFL .446*
Tackles (DE)
For Loss
NFL .511**
Sacks (DE)
NFL .563** -.751*
Fumbles (DE) (S)
Forced
NFL -.688*
Fumbles (S)
Recovered
NFL .456*
Intercep- (CB)
tions
NFL .575**
Passes (CB)
Defensed
!
112!
Limitations, Directions for Future Research
This study is not without issue, and there is also room for further advancement
following the conclusion of it. In measuring only first-round picks in six years of a total
36 of the Draft process having a Combine, this study evaluates a representative sample of
first-round picks, but not NFL players as a whole, with the league containing nearly
1,700 players (Wilco, 2018), only 85 of which were still active from this study in the
league’s nearly 1,700. Other studies similar to this work (Kuzmits & Adams, 2008;
Lyons et al., 2011; Mcgee & Burkett, 2003; Mulholland & Jensen, 2016; Park, 2016
Robbins, 2010; Vincent et al., 2018) have examined full draft classes or positions over
multiple years to gain a larger number of subjects, which leads to avoiding issues like this
study had with tight end, in which only four were selected in the first-round in the six-
year sample. There is value in sampling the way the aforementioned studies did, though
for this study it would have fundamentally changed the basis of it to do so. First-round
picks were chosen because of the great financial investment, at least triple the amount of
the rest of draft picks (Belzer, 2018; Gaines & Yukari, 2017) depending on round, first-
round selections have on their teams. There is not as much risk in mistaking a player in
the final six rounds for a future star as there is in the first round, hence the sample
selection. While the sample could’ve been extended to cover more years to gather more
subjects, reliable collegiate data began to become an issue when the author looked into
including years back to 2002. With interrater reliability, this study has six years of
unquestionable data, but venturing back to the mid- and early-2000s found different
sources stating different statistics for the same player. This was not the case for the
sample in this work, and to put reliable data at risk by contaminating it with factually
!
113!
questionable statistics was a risk deemed fruitless. Should the choice have been made to
use more recent years in which data is reliable and complete, much like the 2007-2012
sample was, the study would run into problems with careers that would be too young to
quantify. Some players peak later in their careers, and should players from the 2013 NFL
Draft and on have been included, there would not have been enough data to fulfill one of
this study’s goals – to show the career-long impact of a player in the NFL. Lyons et al.
(2011) circumvented this road block by studying only four years of a player’s NFL
career, but with the average NFL career down to 2.66 years (Arthur, 2016) and the range
of the shortest careers to longest careers as wide as ever, not taking the entire career of a
player into account leaves incomplete areas and room for oversight. While this study
eliminates six rounds per year, roughly 200 collegiate prospects per season, from
inclusion. While the author believes limiting this study to first-round picks overcomes
some of the root problem in evaluation of prospects by classifying first-round picks as
their own subset of the Draft and evaluating them as the gold standard for this type of
study, there is still some degree of discomfort with a sample size that it was not possible
to expand without putting the data in peril.
In answering Research Question No. 2, struggles came about when attempting to
measure the current Power 5 system against the previous Bowl Championship Series
(BCS) system that was in place at the time of the sample. With conference reformation
taking place through the latter half of the sample and on through the creation of the
College Football Playoff that replaced the BCS in 2014 (Rittenberg, 2014), 19 teams
moved conferences of the 65 that selected a player in the 2007-2012 NFL Drafts. Now
that the dust has settled, to have a strong grasp on the current conference layout and if it
!
114!
has changed competition, diminished or elevated certain conferences, or simply been a
change in affiliation only with no ancillary effects, an update to Research Question No. 2
needs to be done with a newer sample and the College Football Playoff format, rather
than having to retroactively use the Big East as a power conference, as it no longer
sponsors football.
In some cases, prospects did not participate in drills at the Combine, leading the
author to have to turn to the athlete’s Pro Day to collect the data for a given drill. This
was the case for 33 of the 191 prospects studied in this work, most often factoring into
the Vertical Jump measurement. There were also uncollectable data points, 47 of the 573
in the sample, as a select few prospects did not participate in a drill at either their
Combine workout or their Pro Day. These data points were removed from the study and
assigned no value since they did not exist.
Injuries and off-field events, as is the case in sport and life, can change the
trajectory of careers. In this study’s sample, as pointed out in the Results section,
linebackers Keith Rivers, Nick Perry, and Sean Weatherspoon saw varying injuries affect
their careers varying amounts, but certainly enough to affect their statistics measured in
the secondary data. Laron Landry had his career halted early because of multiple
suspensions (Bieler, 2015), quelling a career that would’ve otherwise been near the top of
the safety sample, while Justin Blackmon received an indefinite suspension from the
team that drafted him, the Jacksonville Jaguars for repeated drunken driving arrests
(Breech, 2015). Worst of all, Gaines Adams died of cardiac arrest at the age of 26 just
three years after being the highest-selected defensive end in the sample (Autopsy:
Adams, 2010). Some of these unforeseen circumstances may emerge in interviews with
!
115!
those that suffer from addiction or erratic behavior, and some medical tests may be able
to give an in-depth look at injury susceptibility, but to a certain point, especially in the
case of Adams, there is an element in studies such as this that will not show up in
statistics and can not be controlled for. One possible place to start would be using per-
game measures instead of total statistics, showing a player’s ability when on the field,
though this would lead to a lack of understanding about the statistics he is truly capable
of putting up should he not being able to stay active for the bevvy of issues that can arise.
In the future, expanding the Combine testing battery that is examined could
attempt to gain a more holistic idea of what leads to NFL success, rather than using just
40-yard dash, vertical jump, and bench press. The three-cone drill and shuttle run,
designed to measure agility, and the broad jump, a test of an athlete’s explosion and
lower-body strength (Workouts, 2019) were the three quantifiable physical drills omitted
from this study, but considering the findings of it, to see if any of the three predict NFL
success better than the three correlated in this study would be a worthwhile endeavor.
Similarly with the statistics that were gathered, using a different set of more in-
depth statistics may find different results than what this study did. Using the most basic
of measurables as the author did gives a general understanding of how a player impacted
his team and the game, but Pro Football Focus, NFL Next Gen Stats, StatsLab, amongst
others, are generating statistics far more advanced than the traditional statistical battery.
Defense-Adjusted Value Over Average, Catch Rate, Air Yards Per Target, and Sacks Per
Knockdown are a few of these (Barnwell, 2017), though the author did not have access to
these statistics and therefore were beyond the scope of this study.
!
116!
In measuring only Secondary Data, an outsider’s point of view is taken since no
one directly involved with the NFL and its scouting process is interviewed or surveyed in
this work. One area to further this research in the future would be to include a survey,
focus group, or interview with scouts from NFL teams across the league to share what
they think is most important when evaluating a player. This could lend an extra layer of
expertise to studies of this kind and bring a mixed-method approach that could be viewed
as more reliable than only incorporating statistical analysis of Combine measures v.
collegiate in-game performance.
!
117!
CHAPTER 6
CONCLUSION
In this study, in-game collegiate performance has shown to be a better predictor of
NFL success than NFL Draft Combine drill performance, though across all positions
included in this study, generalizing that statement proves suspect at best. Whether
discussing collegiate in-game or Combine performance as predictors of NFL success
amongst the highly coveted first-round draft picks discussed in this study, the
overarching takeaway from this work is that intricacies are present with every position,
unique skill sets required to have success at each, just as unique evaluation processes are
required in order to confidently determine top talent. This work suggests a new in-game
statistical measuring stick for quarterbacks, running backs, wide receivers, and tight ends
during the evaluation stages in order to create a confident assessment of what a player
can achieve at their position. While in-game collegiate statistics can not be used at the
skill positions on offense, this work shows significant statistical predictive validity on the
defensive side of the ball via statistics amassed by four of the five defensive positions
measured, most substantially defensive tackle and linebacker. In addition, this study
suggests that cornerbacks and offensive linemen can be accurately rated via the
athletically rooted 40-yard dash and vertical jump Combine drills. Outside of those
positional findings, bench press is rejected across all positions as predictive of NFL
success, and other drills are suggested for inclusion in future studies to further this work’s
assertion that the NFL Combine is a brand and media positioning event, rather than one
designed to uniformly determine those that will have success in the NFL. Additionally,
the findings of Research Question No. 2 show that normalization or weighting of data is
!
118!
not required when analyzing Power 5 prospects against non-Power 5 prospects. Further,
the study showed no definitive advantage to attending a Power 5 school, outside of the
safety position, in terms of draft status or future NFL success, with players showing the
ability to succeed from all walks of college football.
!
119!
REFERENCES
Abbasian, R., Sieben, J., Gastauer, A. (2016). Statistical modeling of success in college
and NFL for a star-rated football recruit. International Journal of Statistics and
Applications, Volume 6, Issue 4, pp. 235-240. Retrieved from:
http://article.sapub.org/10.5923.j.statistics.20160604.04.html
Anding, R., Oliver, J. (2015). Football Player Body Composition: Importance Of
Monitoring For Performance And Health. Retrieved from:
https://www.gssiweb.org/sports-science-exchange/article/sse-145-football-player-
body-composition-importance-of-monitoring-for-performance-and-health
Arawolo, O. (2017). Understanding Framing Theory. Retrieved from:
https://www.researchgate.net/publication/317841096_UNDERSTANDING_FRA
MING_THEORY
Arthur, R. (2016). The Shrinking Shelf Life of NFL Players. Retrieved from:
https://www.wsj.com/articles/the-shrinking-shelf-life-of-nfl-players-1456694959
Associated Press. (2007). Michael Vick Sentenced to 23 Months in Jail for Role in
Dogfighting Conspiracy. Retrieved from:
http://www.foxnews.com/story/2007/12/10/michael-vick-sentenced-to-23-
months-in-jail-for-role-in-dogfighting-conspiracy.html
!
120!
Autopsy: Adams died of cardiac arrest. (2010). Retrieved from:
http://www.espn.com/chicago/nfl/news/story?id=4833908
Average playing career length in the National Football League (in years). (2018).
Retrieved from: https://www.statista.com/statistics/240102/average-player-career-
length-in-the-national-football-league/
Balmas, M., Sheafer, T. (2010). Candidate Image in Election Campaigns: Attribute
Agenda Setting, Affective Priming, and Voting Intention. Retrieved from:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1620266
Barnwell, B. (2017). The NFL stats that matter most. Retrieved from:
http://www.espn.com/nfl/story/_/id/20114211/the-nfl-stats-matter-most-2017-
offseason-bill-barnwell
Beckett, K. (2015). Top 15 Worst QBs in NFL History. Retrieved from:
http://www.thesportster.com/football/15-worst-qbs-in-nfl-history/.
Belzer, J. (2018). 2018 NFL Draft First-Round Rookie Salary Projections: What
Mayfield, Barkley, and Darnold Will Make. Retrieved from:
https://www.forbes.com/sites/jasonbelzer/2018/04/27/2018-nfl-draft-1st-round-
rookie-salary-projections-what-mayfield-barkley-and-darnold-will-
make/#6ad2f6a14581
!
121!
Bender, B. (2018). Ranking every NFL quarterback drafted in first round since 2000.
Retrieved from: http://www.sportingnews.com/us/nfl/list/nfl-draft-first-round-
quarterbacks-picks-tebow-newton-manziel-
rodgers/16obdsowna7kk1gnhl9iqrp2pg
Benoit, A. (2014). The safety dance. Retrieved from:
https://www.si.com/2014/03/19/nfl-free-agency-safety-position-evolution
Bieler, D. (2015). LaRon Landry, Fresh Off of 10-Game Suspension, Gets Suspended
Indefinitely. Retrieved from: https://www.washingtonpost.com/news/early-
lead/wp/2015/11/25/laron-landry-fresh-off-of-10-game-suspension-gets-
suspended-indefinitely/?utm_term=.b671abfafb71
Birkett, D. (2017). In NFL draft’s deep cornerback class, size matters. Retrieved from:
https://www.freep.com/story/sports/nfl/lions/2017/03/05/2017-nfl-draft-top-
cornerbacks/98781910/
Blender, C. (2019). Average NFL height and weight by position. Retrieved from:
https://webpages.uidaho.edu/~renaes/251/HON/Student%20PPTs/Avg%20NFL%
20ht%20wt.pdf
Breech, J. (2015). Justin Blackmon faces another DUI charge after 4
th
arrest in five years.
!
122!
Retrieved from: https://www.cbssports.com/nfl/news/justin-blackmon-faces-
another-dui-charge-after-4th-arrest-in-five-years/
Breer, A. (2015). Origins of the Workout Warrior: Mike Mamula changed combine.
Retrieved from:
http://www.nfl.com/combine/story/0ap3000000473741/article/origins-of-the-
workout-warrior-mike-mamula-changed-combine.
Bruce Campbell. (2013). Retrieved from: https://www.pro-football
-reference.com/players/C/CampBr20.htm
Bullock, M. (2015). Differences between the zone and power running schemes. Retrieved
from: https://www.washingtonpost.com/news/football-
insider/wp/2015/05/21/differences-between-the-zone-and-power-running-
schemes/?utm_term=.6bcf47c024ee
Caputo, P. (2010). Caputo: NFL scouting combine can be misleading. Retrieved from:
https://www.theoaklandpress.com/news/caputo-nfl-scouting-combine-can-be-
misleading/article_d68bfaa3-38fc-5756-9084-c22ce82d0af6.html
Charles Rodgers. (2003). Retrieved from: https://www.pro-football
-reference.com/players/R/RogeCh01.htm
!
123!
Clark, K. (2018). Patrick Mahomes II is here to save the deep ball – and destroy the NFL.
Retrieved from: https://www.theringer.com/nfl/2018/9/11/17844496/patrick-
mahomes-ii-is-here-to-save-the-deep-ball-and-destroy-the-nfl
Clark, K. (2018). The NFL’s analytics revolution has arrived. Retrieved from:
https://www.theringer.com/nfl/2018/12/19/18148153/nfl-analytics-revolution
David Carr. (2002). https://www.sports-reference.com/cfb/players/david-carr-1.html.
2002 NFL Combine Results. (2002) Retrieved from:
http://nflcombineresults.com/nflcombinedata.php?year=2002&pos=&college=.
Davis, N. (2017). Ranking NFL Linebackers by Team: Panthers, Broncos Leading the
Way. Retrieved from:
http://www.nfl.com/news/story/0ap3000000931536/article/von-miller-heads-list-
of-best-pass-rushers-ahead-of-2018-season
Demovsky, R. Packers must use free agency again, and GM ‘won’t be afraid’. Retrieved
from: http://www.espn.com/blog/green-bay-packers/post/_/id/46548/packers-
must-use-free-agency-again-and-gm-wont-be-afraid
Dengel, D.R., Bosch, T.A., Burruss, T.P., Fielding, K.A., Engel, B.E., Weir, N.L.,
Weston, T.D. (2013). Body Composition and Bone Mineral Density of National
Football League Players. Journal of Strength and Conditioning Research, Volume
!
124!
28, Issue 1, PP 1-6. Retrieved from:
https://www.ncbi.nlm.nih.gov/pubmed/24149760
Despite Down Year, NFL and Olympics Dominate. (2017). Retrieved from:
http://www.sportsmediawatch.com/2017/01/most-watched-sporting-events-2016-
nfl-olympics-world-series-nba-college-football/.
Dillon, D. (2004). Getting their nose dirty. Retrieved from:
https://web.archive.org/web/20090829075406/http://tsn.sportingnews.com/exclusi
ves/20041011/572815.html
Doll, T. (2013). Some clarification is in order: Average speed by position. Retrieved
from: https://www.milehighreport.com/2013/2/12/3969128/some-clarification-is-
in-order-average-speed-by-position
Duffy, T. (2017). Tom Brady Finds NFL Combine Shirt, Reflects on Proving Scouting
Reports Wrong. Retrieved from: http://bleacherreport.com/articles/2696013-tom-
brady-finds-nfl-combine-shirt-reflects-on-proving-scouting-reports-wrong.
Eckard, W.E. (1998). Is the Bowl Championship Series a Cartel? Some Evidence.
Review of Indutrial Organization, Volume 13. Retrieved from:
https://link.springer.com/article/10.1023/A:1007713802480
!
125!
Eckard, W.E. (2018). Does the NCAA’s Collegiate Model Promote Competitive
Balance? Power-5 Conference Football Versus the NFL. Retrieved from:
http://journals.sagepub.com/doi/10.1177/1527002518798687
[Evolve, IMG]. (2013). ESPN: Jamarcus Russell - Waking Up [Full Feature]. Retrieved
from: https://vimeo.com/64679860.
Fastest 40 Yard Dash Times in NFL History. (2018). Retrieved from:
https://www.statisticbrain.com/fastest-40-yard-dash-times-in-nfl-history/
FBS (I-A) Player Receiving Statistics. (2006). Retrieved from:
https://www.espn.com/college-
football/statistics/player/_/stat/receiving/sort/receivingYards/year/2005.
Feng, E. (2019). The essential guide to predictive college football rankings. Retireved
from: https://thepowerrank.com/guide-cfb-rankings/
Fierro, N. (2015). Scouting Combine evolution underscores thirst for knowledge.
Retrieved from: https://www.mcall.com/sports/football/eagles/mc-scouting-
combine-evolution-0222-20150222-story.html
Football Bowl Subdivision Records. (2017). Retrieved from:
http://fs.ncaa.org/Docs/stats/football_records/2017/FBS.pdf.
!
126!
Fornelli, T. (2017). Friday Five: The greatest Cinderella stories in college football
history. Retrieved from: https://www.cbssports.com/college-football/news/friday-
five-the-greatest-cinderella-stories-in-college-football-history/
Fransen, T. (2017). The NFL scouting combine. Retrieved from:
http://thecrite.com/coloradomesau/nfl-scouting-combine/.
Gabriel, G. (2017). The History and Purpose of the NFL Scouting Combine. Retrieved
from: http://www.profootballweekly.com/2017/02/20/greg-gabriel-the-history-
and-purpose-of-the-nfl-scouting-combine/atyjtvh/.
Gabriel, G. (2018). NFL Draft: Greg Gabriel Explains The Importance of Private QB
Workouts. Retrieved from: https://www.profootballweekly.com/2018/03/09/nfl-
draft-greg-gabriel-explains-the-importance-of-private-qb-workouts/azi38w8/
Gaines, C., Yukari, D. (2017). Here’s how much money players lose when they fall in the
NFL Draft. Retrieved from: https://www.businessinsider.com/nfl-draft-contract-
values-2017-4
Goffman, E. (1974). Frame Analysis: An Essay on the Organization of Experience. New
York, New York. Northeastern University Press.
Gunter, J. (2017). Who Are the 50 Highest-Paid Players For the 2017 NFL Season?
!
127!
Retrieved from:
https://www.cleveland.com/browns/index.ssf/2017/07/ranking_the_50_highest-
paid_pl.html
Harrison, E. (2013). NFL free agency has empowered players but ruined rivalries.
Retrieved from: http://www.nfl.com/news/story/0ap1000000146772/article/nfl-
free-agency-has-empowered-players-but-ruined-rivalries
Hersmeyer J. (2018). For a Passing League, the NFL Still Doesn’t Pass Enough.
Retrieved from: https://fivethirtyeight.com/features/for-a-passing-league-the-nfl-
still-doesnt-pass-enough/
History. (2017). Retrieved from: http://www.nflcombine.net/history/.
Hofheimer, B. (2019). ABC to host two-hour NFL Live special at 2019 NFL Scouting
Combine. Retrieved from: https://espnmediazone.com/us/press-
releases/2019/02/abc-to-host-two-hour-nfl-live-special-at-2019-nfl-scouting-
combine/
Iyer, V. (2017). NFL Combine: 15 greatest performances in Combine history. Retrieved
from: http://www.sportingnews.com/nfl/news/nfl-combine-best-performances-
records-all-time-bo-jackson-deion-sanders-chris-
johnson/1c4qhb1fl0xj31n7ndv5ugvz7h
!
128!
Judge issues warrant for ex-Lions WR Charles Rodgers. (2013). Retrieved from:
https://www.usatoday.com/story/sports/nfl/lions/2013/09/13/judge-issues-
warrant-for-ex-lions-wr-charles-rogers/2810093/
Kania, M. (2016). Re: Do You Think Offensive Linemen Would Officially Get Official
Stats? If so, What Should They Be? [Online Discussion Group]. Retrieved from:
https://www.reddit.com/r/nfl/comments/43a3pw/were_pro_football_reference_a_
website_dedicated/
Kay, A. (2017). 2017 NFL Draft Ratings: Analyzing ESPN’s TV & Live Stream
Performance For Shocking First Round. Retrieved from:
https://www.forbes.com/sites/alexkay/2017/04/28/2017-nfl-draft-ratings-
analyzing-espns-tv-live-stream-performance/#4d3f340f6ed3
Kaylor, J. Here Are the 12 Highest Bench Press Totals in NFL Combine History.
Retrieved from: https://www.cheatsheet.com/entertainment/nfl-patrick-mahomes-
net-worth-rookie-season.html/
Kelly, D. (2015). Not Just Throwing Darts. Retrieved from:
https://www.sbnation.com/nfl/2014/3/6/5473554/2014-nfl-draft-scouting-process-
big-board
!
129!
Kirshner, A. (2016). Jahvid Best ran a 100m Olympic sprint 0.32 seconds slower than
Usain Blot. Retrieved from: https://www.sbnation.com/college-
football/2016/8/13/12465668/jahvid-best-olympics-time-100m-sprint
Knaak, J. (2018). Early days of the Combine. Retrieved from:
https://raiders.exposure.co/early-days-of-the-combine
Kohut, T. (2018). Dallas Cowboys: Will Byron Jones finally have a breakout season?
Retrieved from: https://thelandryhat.com/2018/06/30/dallas-cowboys-byron-
jones-breakout-season/
Kutz, S. (2017). This Chart of the Highest-Paid NFL Players at Every Position is Filled
With Surprises. Retrived from: https://www.marketwatch.com/story/these-are-the-
highest-paid-nfl-players-at-every-position-2017-10-25
Kuzmits, F.E., Adams, A.J. (2008). The NFL Combine: Does It Predict Performance In
The National Football League. Journal of Strength and Conditioning Research.
Volume 22, Issue 6, 1721-1727. http://journals.lww.com/nsca-
jscr/Abstract/2008/11000/The_NFL_Combine__Does_It_Predict_Performance_in
.1.aspx
Legwold, J. (2014). Big DBs in high demand, short supply. Retrieved from:
http://www.espn.com/nfl/draft2014/story/_/page/hotread140416/cornerbacks-
playing-catchup-taller-faster-receivers
!
130!
Lopez, J. (2010). The Rule of 26-27-60 helps predict NFL quarterback success or failure.
Retrieved from: https://www.si.com/more-sports/2010/07/08/qb-rule
Lyons, B.D., Hoffman, B.J., Michel, J.W., Williams, K.J. (2011). On the Predictive
Efficiency of Past Performance and Physical Ability: The Case of the National
Football League. Journal of Human Performance. Volume 24, Issue 2, 158-172.
Retrieved from:
https://www.researchgate.net/publication/233458820_On_the_Predictive_Efficien
cy_of_Past_Performance_and_Physical_Ability_The_Case_of_the_National_Foo
tball_League.
Maisel, I. (2014). Autonomy Set To Benefit Athletes. Retrieved from:
http://www.espn.com/college-football/story/_/id/11321434/autonomy-grants-
power-5-more-control
Manfred, T. (2015). NFL Draft Prospect jumps ridiculously high, historically far at the
combine. Retrieved from: https://www.businessinsider.com/byron-jones-nfl-
combine-results-2015-2
Mathews, E.M., Wagner, D.R. (2010). Prevalence of Overweight and Obesity in
Collegiate American Football Players, By Position. Journal of American College
Health. Volume 57, Issue 1. Retrieved from:
!
131!
https://www.researchgate.net/publication/23151333_Prevalence_of_Overweight_
and_Obesity_in_Collegiate_American_Football_Players_by_Position
Matt Jones. (2009). Retrieved from: https://www.pro-football
-reference.com/players/J/JoneMa00.htm.
Mays, R. (2017). The fading art of the deep ball. Retrieved from:
https://www.theringer.com/2017/8/1/16095332/nfl-deep-passing-offense-
inefficiency-week-78527887e6bf
McCaffrey, D. (2015). The Importance of the NFL Draft. Retrieved from:
http://lastwordonsports.com/2015/04/11/the-importance-of-the-nfl-draft/
Mcgee, K.J., Burkett, L.N. (2003). The National Football League Combine: A Reliable
Predictor of Draft Status? Journal of Strength and Conditioning Research,
Volume 17, pp. 6-11. Retrieved from:
https://pdfs.semanticscholar.org/8628/bec98a3c5218815233c8b650b4ff988046e1.
pdf
McGinest, W. (2018). Von Miller Heads List of Best Pass Rushers Ahead of 2018
Season. Retrieved from:
http://www.nfl.com/news/story/0ap3000000931536/article/von-miller-heads-list-
of-best-pass-rushers-ahead-of-2018-season
!
132!
McCombs, M., Llamas, J.P., Lopez-Escobar, E., Rey, F. (1997). Candidate Images in
Spanish Elections: Second-Level Agenda-Setting. Journal of Mass
Communication Quarterly. Volume 74, Issue 4, 703-717.
http://www.aejmc.org/home/wp-content/uploads/2012/09/Journalism-Mass-
Communication-Quarterly1997-McCombs-703-17.pdf
Microsoft Attention Spans Research Report. (2015). Retrieved from:
https://www.scribd.com/document/265348695/Microsoft-Attention-Spans-
Research-Report
More Ratings: NFL Combine, MLB, Gymnastics, NASCAR Xfinity. (2018).
http://www.sportsmediawatch.com/2018/03/nfl-combine-ratings-spring-training-
gymnastics/
Mulholland, J, Jensen, S.T. (2014). Predicting The Draft and Career Success of Tight
Ends in the National Football League. Journal of Quantitative Analysis in Sports.
Volume 10. Issue 4. PP. 381-396. Retrieved from:
https://www.degruyter.com/view/j/jqas.2014.10.issue-4/jqas-2013-0134/jqas-
2013-0134.xml
Mulholland, J., Jensen, S.T. (2016). Projecting the Draft and NFL Performance of Wide
!
133!
Receiver and Tight End Prospects. Retrieved from:
http://chance.amstat.org/2016/11/draft-and-nfl-performance/.
National Scouting Combine. (2018). Retrieved from: https://operations.nfl.com/the
-players/getting-into-the-game/national-scouting-combine/
Niesen, J. (2018). The computer poll uprising: Creators of the BCS’s most controversial
components. Retrieved from: https://www.si.com/college-
football/2018/07/11/bcs-computer-rankings-polls-formula-sagarin-billingsley
[NFL]. (2015). Byron Jones 12’3” Broad Jump Sets World Record 2015 NFL Combine.
Retrieved from: https://www.youtube.com/watch?v=n0UeHxglMJ4
[NFL]. (2016). Jalen Ramsey (Florida St., DB) 2016 NFL Combine Highlights. Retrieved
from: https://www.youtube.com/watch?v=-ZibOYmRNWU
NFL announces free fan opportunities for NFL Combine. (2016).
http://www.nfl.com/news/story/0ap3000000760218/article/nfl-announces-free-
fan-opportunities-for-scouting-combine.
NFL Sacked Single-Season Leaders. (2017). Retrieved from: https://www.pro-football
-reference.com/leaders/pass_sacked_single_season.htm
!
134!
Ourand, J. (2017). NFL Network Sees Big Jump In Scouting Combine Viewers On TV,
Online. Retrieved from:
https://www.sportsbusinessdaily.com/Daily/Issues/2017/03/09/Media/NFL-
Network.aspx
Passing. (2018). Retrieved from:
https://nextgenstats.nfl.com/stats/passing#average-time-to-throw
Park, P. (2016). Does the NFL Combine Really Matter. Retrieved from:
https://www.stat.berkeley.edu/~aldous/Research/Ugrad/Paul_Park.pdf.
Patsko, S. (2016). How each team has – or has not – built through the draft. Retrieved
from:
http://www.cleveland.com/browns/index.ssf/2016/04/how_each_nfl_team_has_-
_or_has.html
Porter, R. (2016). The 100 most-watched TV programs of 2016: Super Bowl 50 leads by a
mile. Retrieved from: http://tvbythenumbers.zap2it.com/more-tv-news/the-100-
most-watched-tv-programs-of-2016-super-bowl-50-leads-by-a-mile/.
Ranking NFL 4-3 defensive lines by weight: Giants are the heaviest, Cowboys the
!
135!
lightest. Retrieved from:
https://www.bloggingtheboys.com/2015/8/12/9129785/ranking-nfl-4-3-defensive-
lines-by-weight-giants-heaviest-cowboys-lightest
Reiss, M. (2016). Most notable trades among Bill Belichick’s 121 with Patriots.
Retrieved from: http://www.espn.com/blog/new-england-
patriots/post/_/id/4797150/most-notable-trades-among-bill-belichicks-121-with-
patriots
Rittenberg, A. (2014). How the playoff came to be. Retrieved from:
http://www.espn.com/college-football/story/_/id/12002638/an-oral-history-
college-football-playoff
Robbins, D.W. (2010). The National Football League (NFL) Combine: Does Normalized
Data Better Predict Performance in the NFL Draft? Journal of Strength and
Conditioning Research, Volume 24, Issue 11, 2888-2899. Retrieved from:
http://journals.lww.com/nsca-
jscr/pages/articleviewer.aspx?year=2010&issue=11000&article=00002&type=abs
tract
Rosenthal, G. (2012). The Best First-Round RBs of The Last Decade. Retrieved from:
http://www.nfl.com/news/story/09000d5d82857cee/article/the-best-firstround-rbs-
of-the-last-decade
!
136!
Rosenthal, G. (2013). All-Pro Team Headlined by Adrian Peterson, J.J. Watt. Retrieved
from: http://www.nfl.com/news/story/0ap1000000125147/article/allpro-team-
headlined-by-adrian-peterson-jj-watt
Ruiz, S. (2016). 10 players who proved how little the NFL combine matters. Retrieved
from: http://ftw.usatoday.com/2016/02/nfl-combine-best-worst-performance-
busts-tom-brady-40-time.
Sagarin, J. (2005). Jeff Sagarin Ratings. Retrieved from:
https://www.usatoday.com/sports/ncaaf/sagarin/2004/team/
Sando, M. (2007). 4-3 teams emphasize speed, not size, at linebacker. Retrieved from:
http://www.espn.com/nfl/columns/story?columnist=sando_mike&id=2995887
Saturday’s NFL Network Scouting Combine viewership up 91% to 529,000. (2015).
Retrieved from: https://sportstvratings.com/saturdays-nfl-network-scouting-
combine-viewership-up-91-to/1520/
Scouts Inc. (2009). Scouts Inc. on defensive tackles. Retrieved from:
http://www.espn.com/college-sports/recruiting/football/news/story?id=2350477
Secora, C.A., Latin, R.W., Berg, K.E., Noble, J.M. (2004). Comparison of Physical and
!
137!
Performance Characteristics of NCAA Division I Football Players: 1987 and
2000. Journal of Strength and Conditioning Research. Volume 18. Issue 2. PP:
286-291. Retrieved from:
https://nebraska.pure.elsevier.com/en/publications/comparison-of-physical-and-
performance-characteristics-of-ncaa-di
Shook, N. (2017). Inside look at the NFL’s most popular run blocking concepts.
Retrieved from:
http://www.nfl.com/news/story/0ap3000000794734/article/inside-look-at-the-
nfls-most-popular-run-blocking-concepts
Silverman, S. (2012). The NFL Combine is Overrated. Retrieved from:
http://chicago.cbslocal.com/2012/02/21/silverman-nfl-combine-is-overrated/
Smith, M.D. (2018). Seven first-team All-Pros were snubbed by the Pro Bowl. Retrieved
from: https://profootballtalk.nbcsports.com/2018/01/05/four-first-team-all-pros-
were-snubbed-by-the-pro-bowl/
Sobleski, B. (2016). Not to be overlooked: A pass-rusher’s natural advantage to being
short. Retrieved from: https://bleacherreport.com/articles/2669199-not-to-be-
overlooked-a-pass-rushers-natural-advantage-to-being-short
Staples, A. (2018). The Chaos and Consequences of the BCS, 20 Years After Its
!
138!
Inaugural Season. Retrieved from: https://www.si.com/college-
football/2018/07/09/bcs-history-20th-anniversary-controversy-tennessee-florida-
state
Staff, TIME. (2010). 50 Best Websites 2010. Retrieved from:
http://content.time.com/time/specials/packages/article/0,28804,2012721_2012880
_2012751,00.html
Tanier, M. (2005). The 4-3 vs. the 3-4. Retrieved from:
https://www.footballoutsiders.com/strategy-minicamps/2005/4-3-vs-3-4
Teicher, A. (2018). Does Eric Berry have another comeback in him? Don’t bet against it.
Retrieved from: http://www.espn.com/blog/kansas-city-
chiefs/post/_/id/24881/does-eric-berry-have-another-comeback-in-him-dont-bet-
against-it
The history of NFL Free Agency. (2018). Retrieved from:
https://frontofficenfl.com/2018/03/13/the-history-of-nfl-free-agency/
Top 15 Most Popular Sports Websites. (2018). Retrieved from:
http://www.ebizmba.com/articles/sports-websites.
Tracy, M. (2016). A.P. Poll about to take a back seat to College Football Playoff
!
139!
rankings. Retrieved from:
https://www.nytimes.com/2016/11/01/sports/ncaafootball/ap-poll-college-
football-playoff-rankings.html
Travis, C. (2008). This Vince Young Melt Down Is Getting Uglier…and Scary. Retrieved
from: https://deadspin.com/5047255/this-vince-young-melt-down-is-getting-
uglierand-scary
Troy Williamson. 2010. Retrieved from: https://www.pro-football
-reference.com/players/W/WillTr01.htm
Tversky, A., Kahneman, D. (1991). Loss Aversion in Riskless Choice: A Reference
Dependent Model. Quarterly Journal of Economics. Volume 106, Issue 4, 1039-
1061. Retrieved from:
https://www.jstor.org/stable/2937956?seq=1#page_scan_tab_contents
Tucker, C.J. (2017). Charles Rogers says pain pills, injuries ruined career, not
marijuana. Retrieved from:
http://www.freep.com/story/sports/nfl/lions/2017/04/19/charles-rogers-says-pain-
pills-injuries-ruined-career-not-marijuana/100682576/
Vincent, L.M., Blissmer, B.J., Hatfield, D.L. (2018). National Scouting Combine Scores
!
140!
As Performance Predictors In The National Football League. Journal of Strength
and Conditioning Research. Volume 33. Issue 1. PP 104-111. Retrieved from:
https://journals.lww.com/nsca-
jscr/Fulltext/2019/01000/National_Scouting_Combine_Scores_as_Performance.1
2.aspx
Wertheim, J. (2011). The Man Who Isn’t There. Retrieved from:
https://www.si.com/vault/2011/10/31/106125502/the-man-who-isnt-there.
What goes on at the Combine. (2019). Retrieved from:
http://www.nfl.com/combine/workouts
What The Stats Tell Us About Drafting Positions By Round. (2015). Retrieved from:
https://www.arrowheadpride.com/2015/2/20/8072877/what-the-statistics-tell-us-
about-the-draft-by-round
White, D. 2017, March. John Ross 40 yard dash.
https://www.youtube.com/watch?v=eI6YVs8mvvA
Wilco, D. (2018). Colleges most represented on 2018 NFL rosters. Retrieved from:
https://www.ncaa.com/news/football/article/2018-09-11/colleges-most-
represented-2018-nfl-rosters
!
141!
Wood, S. (2004). NFL opens combine to curious cameras. Retrieved from:
https://usatoday30.usatoday.com/sports/football/nfl/2004-02-18-nfl-open-
combine_x.htm.
Workouts. (2019). Workouts. Retrieved from: http://www.nfl.com/combine/workouts
!
142!
VITA
MIKE GALLAGHER
Education: B.A. Mass Communication. Augsburg University,
Minneapolis, Minnesota 2012
M.A. Brand and Media Strategy. East Tennessee State
University, Johnson City, Tennessee 2019
Professional Experience: Intern, IHeart Media, KFAN FM 100.3, Minneapolis,
Minnesota 2011-2013
Creator/Producer/On-Air Host/Play-By-Play Voice,
Minnesota Intercollegiate Athletic Conference,
Minneapolis, Minnesota 2011-2017
Play-By-Play, Analyst, Sideline Reporter, Master of
Ceremonies, MG Media, Minneapolis, Minnesota
2011-2017
Sports Producer/Manager, Distribution and Partnerships,
Bring Me The News, Minneapolis, Minnesota 2012-
2014
Producer/On-Air Host, Minnesota Broadcasters
Association, Minneapolis, Minnesota 2014-2017
Academic Advising Coordinator, Walden University,
Minneapolis, Minnesota 2014-2017
Play-By-Play/Media Relations, Elizabethton Twins,
!
143!
Elizabethton, Tennessee 2017-2018
Graduate Assistant/Network Media Assistant, East
Tennessee State University, Johnson City,
Tennessee 2017-2019
Educational Honors: Outstanding Thesis Award, Department of Media
and Communication, East Tennessee State
University 2019
