43
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Examining the Link Between
Poker Room Business Volume and
Gaming Activity in Slot and Table
Games: A Closer Look at a Key
Assumption in the Full Service
Theory
Anthony F. Lucas, Ph.D.
Abstract
Results from three different Nevada hotel-casinos failed to support the popular
notion that poker rooms drive business to the slot and table game areas of the casino
oor. This result not only questions the validity of a key and somewhat bold operating
assumption, it casts a shadow of doubt on the broader Full Service Theory, as applied to
the casino oor. Additionally, this work extends Ollstein (2006) by empirically examining
the relationships between the daily business volumes of poker rooms and both critical
gaming centers (i.e., slots and table games). Five of six key results question the wisdom
of offering live poker, based on the assumption of indirect revenue contribution to
slots and table games. Double-log time series models are advanced to analyze the daily
operating results of three casinos over a seven-month period, offering a rare and insightful
look at actual casino performance data.
Keywords: Poker room operations, casino management, slot operations, table game
operations, operations analysis
Introduction
There is no shortage of claims that poker rooms drive business to key casino prot
centers such as slots and table games (Cosgrove-Mather, 2005; Grochowski, 2005;
Legato, 2010; McGowan, 2010; Taucer, 2004; Walters, 2003; Wiser, 2004). However,
there is a shortage of compelling empirical support for these claims. Only one published
study has examined the relationship between the business volumes of the poker room
and the slot oor (Ollstein, 2006). Worse yet, no studies have examined the relationship
between the business volumes of poker rooms and table games.
Poker rooms are somewhat notorious for their inability to produce competitive levels
of prot per square foot (Grochowski, 2005; McGowan, 2010; Taucer, 2005). Others
have described live poker as something of a loss-leader (McGowan 2010; Grochowski
2005), existing only because of assumed revenue contributions to slots and table games.
This assumption of contributions to other gaming areas is absolutely critical. For some,
Anthony F. Lucas, Ph.D.
Professor
William F. Harrah College of
Hotel Administration
University of Nevada, Las
Vegas
Email:
44
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
insufcient prots from operations combined with the weakly supported assumption of
external revenue contributions makes the poker room worthy of a closer look.
Although there are Las Vegas poker rooms that continue to draw crowds, several
operators have begun to question the wisdom of offering live poker, in spite of the
widely held assumption of external revenue contributions. For example, in 2011,
casino executives at the Gold Coast closed the poker room, using the oor space to
accommodate additional slot machines (Sieroty, 2011). In 2008, the Las Vegas Hard
Rock constructed an extravagant $30-million, 18-table poker lounge (Boncek, 2008).
By 2010, the poker lounge was closed, with live poker relegated to a far less glamorous
location that accommodated only eight games (Arnett, 2010). This downsizing occurred
following a major hotel and casino expansion, suggesting the decision was not driven
by a shortage of gaming space. Other notable poker room closures included the Paris
Hotel Casino in 2008, and the Tropicana, Aliante Station, and Silverton, all three of
which were shuttered in 2012 (Mehaffey, 2012). While live poker rebounded from near
extinction in the late nineties to reach unimaginable heights in the mid-aughts, it would
seem as though some are once again beginning to wonder whether it is the best use of
casino oor space.
An improved understanding of the poker room’s contribution to the casino is a
particularly important issue for operators who (1) manage space-constrained casinos
with viable alternative uses for the poker room’s oor space and (2) are considering the
addition of a poker room to an existing property. Developers would also benet from
such knowledge, as they must decide whether to include a poker room in the plans for
a casino expansion or an entirely new property. These operators and developers cannot
afford to assume that poker rooms supply signicant revenue contributions to slots and
table games, based on little more than popular opinion.
Casino executives are pressured to offer the game mix that optimizes prots, while
simultaneously complying with a host of curious operating paradigms such as the full
service theory. Among other things, the full service theory holds that the presence of
the poker room increases wagering volume on both table games and slots. This study
empirically examines the relationships between the business volume of the poker room
and those of both slots and table games, providing a detailed examination of the poker
room’s role in the full service theory. The results will help gaming managers and casino
developers better understand a critical operating assumption that affects both important
game mix decisions and overall casino prots.
Literature Review
Direct and Indirect Revenues
It will be helpful to dene and few key terms before an in-depth review of the
literature. In a hotel-casino resort, “non-gaming amenities” refers to on-site facilities
such as restaurants, retail outlets, bars, and the hotel itself. This is not an exhaustive list,
as there are many forms of non-gaming amenities. While a limited form of gaming can
occur in a non-gaming amenity, such as keno service within a restaurant, non-gaming
amenities typically do not feature gaming.
“Gaming amenities” informally refers to prot centers within the overall casino
which produce marginal operating prots, if any at all. The term “amenity” is assigned
to these gaming areas because the amount of operating prot they produce is nominal
in comparison to that generated by the critical casino prot centers such as slots. For
example, bingo and keno are often referred to as gaming amenities, or ancillary gaming
45
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
activities. Poker rooms and race and sports books also fall into this category for many
operators.
It is also important to dene the difference between direct and indirect revenue
contributions, a distinction advanced by early gaming scholars (Dandurand and
Ralenkotter, 1985; Roehl, 1996). For example, direct revenues in a casino-operated
restaurant stem from purchases of items such as meals, which are sold in the outlet itself.
Additionally, these restaurants are often assumed to produce indirect revenues in other
areas of the resort by either attracting patrons to the property or keeping them from
leaving (Brock, Newman and Thompson, 1992). Many operators assume that restaurants
draw diners to the resort who eventually nd their way into the casino (Lucas and Santos,
2003). Once in the casino, these diners produce gaming win and ultimately, operating
prots for the casino. The gaming revenues generated by diners who were supposedly
lured to the resort by its restaurants would be considered a form of indirect revenue
produced by or at least informally credited to the restaurants.
The Full Service Theory
This may be more of an operating assumption than a theory, at least from an
academic perspective. In any case, it is quite popular among gaming industry executives.
Simply put, the full service theory holds that amenities attract play that would otherwise
be absent (Lucas and Kilby, 2008). It applies to both non-gaming and gaming amenities.
For example, by having a poker room (a gaming amenity), it is assumed that a casino
becomes increasingly attractive to groups of two or more people containing at least one
person interested in playing poker. In general, the full service theory is based on the idea
of capturing more customers by casting a wider net. The width of this “net” is dened by
the diversity of the games offered by the casino.
In practice, the full service theory is often used to justify the existence of an amenity
which is failing to produce sufcient direct prots. The justication usually takes the
following form: While it may be obvious the amenity is unable to demonstrate acceptable
operating prots, eliminating it would cause a decline in critical gaming volumes such
as slot and table game play. It is important to note here that revenues and expenses from
poker rooms are almost never reported on the table game department’s internal income
statement. A separate income statement is prepared, including only those revenues
and expenses produced by the poker room. Alternatively stated, poker tables are not
considered table games, which may be counterintuitive to those outside of the gaming
industry.
The full service theory defense is most likely to be invoked by managers of the
underperforming department when there are viable alternative uses of their oor space.
However, measuring an ailing amenity’s indirect revenue contributions is very difcult
and remains a challenge for most in the gaming industry (Lucas and Kilby, 2008). When
the claims of indirect revenues cannot be denitively established or refuted, the manager
of the amenity is able to play this troubling reality to something of an inconclusive
stalemate. Should the decision makers (e.g., senior management) subscribe to the widely
accepted full service theory, the ailing amenity is likely to dodge elimination. Poker room
managers are among those who are naturally familiar with this game.
Figure 1 is offered as an overview of the full service theory. It is based on
descriptions found in Lucas and Kilby (2008) who discuss but do not endorse the theory.
The contents of Figure 1 are further described in the ensuing paragraphs.
46
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Figure 1. Illustration of causal assumptions included in the Full Service Theory, as
applied to hotel-casino resorts.
Starting at the far left of Figure 1, non-gaming amenities are shown to inuence
casino business volumes. While amenities that produce substantial direct prots such
as hotels and nightclubs do not have to rely on the full service theory for survival, the
overall performance of others such as showrooms and even restaurants is sometimes
questioned (Lucas & Kilby, 2008). In this case, overall performance is dened as the
sum of direct prots and all estimated or assumed indirect prot contributions. The
full service theory is often used to justify annual operating losses in casino-operated
restaurants and showrooms.
Couched within the broader framework of the full service theory, the focus of the
current study resides within the Casino block of Figure 1. This area is further divided into
two sections, Gaming Amenities and Primary Gaming Prot Centers. Within the Gaming
Amenities section, this study empirically examines the relationships illustrated by the
bold line linking Poker Rooms to both Slot Play and Table Games Play. When direct
prots in the poker room are deemed unacceptable, continued operation often relies on
the assumption that the poker room drives business to critical gaming areas such as slots
and table games (as illustrated in Figure 1). Alternatively stated, if casino executives feel
that the poker room is not producing enough operating prot from live poker, they must
justify the existence of the poker room based on the belief that it is driving business in
other key areas of the casino.
Although not shown in Figure 1, the full service theory extends to the game level.
For example, there is much debate regarding the appropriate game mix within the table
game operation. Specically, management strives to optimize operating prots by
managing the supply mix, in terms of betting limits and the types of games offered (e.g.,
blackjack, roulette, craps, baccarat, etc.).
Supply Mix
The Casino section of Figure 1 illustrates something of a supply/game mix
problem occurring within the space of the entire casino oor. Broadly speaking, space
Restaurants
Bars & Nightclubs
Hotels
Showrooms & Cabarets
Retail Outlets
Multipurpose Arenas
Recreational Amenities
(e.g., Rollercoasters,
Bowling Alleys, Movie
Theaters, etc.)
Children’s Recreation
Centers
Other (e.g., Art Galleries,
Exhibitions, and more)
Non-Gaming Amenities Casino
Slot Play
Table Games
Play
Gaming Amenities Critical Gaming Activities
Primary Gaming
Profit Centers
Table Games
Slots
Race & Sports Books
Poker Rooms
Bingo Rooms
Keno Lounges
Other (i.e., other
than those listed
above, slots, and
table games)
Spas & Salons
47
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
optimization challenges are abundant in the hospitality sector and have been for many
years. Most of these challenges have been addressed by way of revenue management
techniques within the hotel, airline, and restaurant industries (Hanks, Noland & Cross,
1992; Smith, Leimkuhler & Darrow, 1992; Kimes & Thompson, 2004). While few
would argue that an optimal supply mix leads to optimal revenues, the gaming question
illustrated in Figure 1 poses an additional hurdle.
To apply revenue management techniques, management must be able to forecast
and measure “total” revenues. However, claims of indirect revenues confuse the issue.
Specically, how can a revenue management process be applied to an area in which
“total” revenues cannot be directly computed? Again, in the case of the poker room, total
revenues would equal direct revenues plus any indirect revenues credited to the poker
room. Without a sophisticated measurement process, these indirect revenues can only
be crudely estimated. A mere claim of indirect revenues gives the poker room manager
grounds to question the accuracy of any optimization process that does not include this
elusive estimate.
Spillover Effect
From the retail literature, the spillover effect is consistent with the general idea of the
full service theory. The spillover effect is the term used to describe the condition whereby
the sales of one store are found to affect those of another. It is often attributed to anchor
stores in retail shopping centers (Eppli and Schilling, 1995). In the current study, the key
issue is whether the poker room attracts gamblers who spillover into other areas of the
casino.
Cherry Pickers
The idea of cherry picking also comes from the retail literature. Contrary to the full
service theory, cherry pickers have specic targets in mind. For example, cherry picking
is thought to occur when stores aggressively drop prices on selected items and experience
no increase in the sales of complementary items. In retail settings, even loss-leader
pricing has failed to increase the sales of the full-priced complementary goods (Walters
and Rinne, 1986; Walters and MacKenzie 1988). For example, if hotdogs were priced
below cost, no increase would occur in the sales of complimentary items such as hotdog
buns, ketchup, and mustard. Only increases in the purchases of the loss-leader item
would occur (i.e., the hotdogs).
Within the gaming industry, loss-leader pricing strategies in casino-operated
restaurants have failed to produce corresponding increases in the wagering activity
of key gaming areas (Lucas and Brewer, 2001). Cherry picking was advanced as one
explanation for this disappointing result. That is, the loss-leader pricing was thought
to have attracted deal-prone restaurant customers to the property, with little interest in
gaming, if any at all.
Regarding the current study, it is certainly possible that poker players are only
interested in playing poker. Poker is a notoriously slow game that usually features a very
low cost per hour for the player (Taucer, 2005). It is also somewhat unique in that players
are pitted against each other and compete for each other’s bankroll. At a minimum, a
subdued form of cherry picking by poker players is a possibility worthy of consideration.
48
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Trade Literature
A trade literature review is important to the extent that it establishes the widespread
acceptance of some critical assumptions related to the current study. First, while there
are always exceptions, poker rooms are not known for producing stellar operating prots
(Cosgrove-Mather, 2005; Gellar, 2009). Some have gone as far as comparing poker
rooms to loss-leader pricing strategies (McGowan, 2010; Grochowski, 2005). Second,
many operators and industry pundits believe that poker rooms drive slot and table
game play, which is actually a causal statement (Cosgrove-Mather, 2005; Grochowski,
2005; Legato, 2010; McGowan, 2010; Taucer, 2004; Walters, 2003; Wiser, 2004).
This assumption is critical to the existence of the poker rooms, especially when direct
operating prots are in short supply. The assumed indirect revenue contributions are
thought to occur from two sources. The rst origin is crossover play, which is based on
the assumption that poker players produce meaningful gaming activity in other areas of
the casino (Byrne, 2010). The second source is similar to what Lucas & Kilby (2008)
describe as the entourage effect. In this case, it is assumed that other parties accompany
poker players to the casino, and these other parties engage in meaningful gaming activity
outside of the poker room (Cosgrove-Mather, 2005; Wolf, 2010).
Claims such as the ones described in the previous paragraph are the bedrock of
the full service theory. Similar if not identical claims are made about bingo and race
and sports books. The next section reviews the extant literature related to the alleged
relationships described within the Casino block of the full service theory, as depicted in
Figure 1.
Indirect Contributions of Gaming Amenities
Only Ollstein (2006) has examined the link between the poker room and slot
play, as illustrated in Figure 1. He examined daily performance data ranging from
February 1, 2005 to August 31, 2005, in an effort to assess the nature of the relationship
between daily poker room rake and aggregate slot coin-in. It is important to note that
Ollstein’s data were collected at a Las Vegas Strip resort during a time that is generally
considered to be the zenith of live poker’s popularity (i.e., c. 2005 – 2006).
Ollstein found a signicant and positive relationship between daily poker room
rake and daily aggregate coin-in (B = 98.63; p < 0.05). This result indicated that a one-
dollar increase in poker room rake could be expected to produce a $98.63 increase in slot
wagers. The rake variable represented the aggregate dollar amount of daily fees collected
from poker players. Other than hourly poker room headcount data, which most casinos
do not have, rake is considered to be the best available business volume indicator. On the
slot side, the coin-in variable represented the aggregate daily dollar amount of wagers
accepted in coin- or voucher-operated wagering devices. Ollstein did not examine the
relationship between poker room rake and table game drop, as illustrated in Figure 1.
Like the current study, Ollstein (2006) analyzed times series data using a model
consisting of the following types of predictor variables: Day of the week, holiday periods,
rake, property-wide promotions, special events, and ARMA terms. The ARMA terms
were used to create an independent error process. While it may seem simplistic, Ollstein’s
model explained 89% of the daily variation in the resort’s daily coin-in.
In spite of the positive relationship, Ollstein expressed concern for his result. First,
he noted that the casino could only expect to retain 7.5% of the expected $98.63 increase,
as the regression coefcient represented wagering volume and not expected win. Second,
he mentioned the incremental operating costs associated with processing the additional
49
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
wagers. Third, after comparing estimates of what the poker room oor space could
produce as a slot area, he questioned whether the magnitude of the regression coefcient
was sufcient to sustain poker room operations.
Staying within the Casino block of Figure 1, Lucas and Brewer (2001) found a
positive relationship between the aggregate daily bingo headcount and aggregate daily
coin-in. However, more recent research employing the same methodology has produced
different results. Lucas, Dunn, and Kharitonova (2006) failed to nd a statistically
signicant relationship between daily bingo headcount and coin-in levels. This result
was produced using data from both a Las Vegas resort and a Southern California Indian
casino.
Abarbanel, Lucas, and Singh (2011) is the only study to have empirically examined
the relationship between race and sports book business volume indicators and daily coin-
in levels. Their original model included predictor variables representing the daily gaming
activity in both the race and sports books, but neither variable produced a statistically
signicant effect on the criterion variable - daily aggregate coin-in. The theoretical
model was well specied in Abarbanel et al., predicting 90% of the daily variation in the
dependent variable, over a 250-day period ranging from January 1, 2009 to September 7,
2009.
Abarbanel et al. (2011) included commentary from operators and industry pundits
claiming the existence of a positive relationship between book and casino gaming levels
(e.g., Manteris, 1993). Manteris cited the importance of the sports book’s indirect revenue
contributions; given its less than noteworthy direct revenues. In Eng (2008), Manteris
cited the draw power of a state-of-the-art race and sports book, again claiming indirect
contributions to other areas of the resort. This is further evidence of subscription to the
full service theory. However, with regard to slot play, the results produced in Abarbanel et
al. failed to support Manteris’ claims.
Time Series Models in Gaming
The current study employed the base model described in Ollstein (2006), which
rst appeared in Lucas and Brewer (2001). This model has been an effective predictor of
both daily aggregate coin-in (Abarbanel et al., 2011; Ollstein, 2006; Suh, 2006) and daily
aggregate table game drop (Lucas, 2004; Lucas 2010; Suh, 2006), making it appropriate
for this study. In fact, the R
2
values for the cited coin-in models ranged from a low of
89% to a high of 91%. The R
2
values for the cited table game drop models ranged from a
low of 70% to a high of 91%. Five of the six drop models produced an R
2
value of least
90%, only the sixth model generated an R
2
of 70%. The sixth model produced its result
from data donated by the management of a high-end Las Vegas Strip resort. Because of
its premium clientele, this casino experienced much more variance in its daily table game
drop, which was caused by marker (i.e., credit) transactions. This is not an unusual result/
condition for casinos that cater to premium table game players (a.k.a. high-rollers).
One minor difference from the coin-in model described in Ollstein (2006) was the
addition of a linear trend variable, which appeared in the coin-in model advanced in
Lucas et al. (2006). Figure 2 illustrates the specic theoretical model tested in the current
study. The operalization of each model variable is described in the Method section.
50
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Figure 2. Theoretical model designed to predict daily coin-in and table game drop.
Hypotheses
With regard to the Casino block of Figure 1, ve of the hypotheses tested by previous
researchers failed to support the full service theory (Abarbanel et al., 2011; Lucas, et
al., 2006). However, two of the hypothesis tests provided at least partial support for the
theory (Lucas and Brewer, 2001; Ollstein, 2006). Given the low number of hypotheses
tested and the mixed results of previous researchers, a directional hypothesis was not
advanced. Instead, the following null hypothesis was tested using data from three
different hotel-casino resorts.
H
0
: B
Rake
= 0
In the above null hypothesis “B
Rake
” represented the regression coefcient for the
poker room rake variable in each of the models tested (i.e., coin-in and drop models),
for each of the three donor properties. All hypothesis tests were conducted within the
framework of the model depicted in Figure 2.
Daily Poker Room
Rake
Days of the Week
Holidays
Special Event Days
Linear Trend
ARMA Terms
(as needed)
Daily Coin-in
(Slots)
Daily Drop
(Tables)
51
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Methodology
Data Sources
Audited secondary data were donated by three different Las Vegas hotel-casino
resorts. The donor properties are referred to as Resorts 1, 2, and 3, as the management
teams requested that the actual names of the properties not appear in the study. All three
properties have consistently generated in excess of $72 million in annual gross gaming
win, placing them in the top tier of the Nevada Gaming Control Board’s revenue reporting
hierarchy. The number of tables in the poker rooms varied by resort. At the time the data
were collected, Resort 1 operated only 8 games, while Resorts 2 and 3 operated 22 and 12
games, respectively.
Resort 1 was located on the Las Vegas Strip, featured over 3,500 hotel rooms, and
catered to customers attracted to what its management described as mid to low-end price
points. Resort 2 also resided on the Las Vegas Strip, offered over 3,500 hotel rooms, but
catered to what its management referred to as a mid-level to high-end clientele. Because
of its presence in the premium player market, Resort 2 also offered an impressive array
of restaurant and entertainment amenities. Resort 3 was located off the Las Vegas Strip,
featured less than 750 hotel rooms, and the casino catered primarily to the Las Vegas-area
residents. As the off-Strip location would suggest, Resort 3 offered much less in terms of
restaurant and entertainment amenities and operated most facets of the property at price
points well below those of the Strip casinos.
Sample Period
Each resort’s data were sequentially ordered by day, with the 217-day samples
beginning on February 3, 2009 and ending on September 7, 2009. These were the dates
for which all three properties could supply a common data set. However, this general
sample period is quite common in gaming studies of Las Vegas resorts, as it does
not include the typical trough and peak business periods. For example, November is
generally a big convention month, which lls hotel rooms with low-gaming-value guests.
Additionally, the end of November includes Thanksgiving, the beginning of the holiday
season, which extends to the end of December. This seven- to eight-week-long trough
in casino business is followed by some of the busiest gaming days of the year – New
Year’s Eve and New Year’s Day. This peak period may extend for a week, which will
be followed by a sharp decline in daily gaming levels. It is very difcult for techniques
such as time series regression analysis to produce a set of coefcients which are able to
incorporate the consecutive phenomena of the trough period, peak period, and subsequent
fall-off in daily gaming volume, while remaining sensitive to the changes occurring in
the relatively stabile periods of the year. Further, given the current form of the time series
model, the inclusion of additional binary variables representing the extended trough
period and post-holiday decline would greatly increase the level of multicollinearity. Of
course, this condition often obscures a clean look at the key variable, which in this case
is poker rake. In fact, none the studies reviewed in this article included the months of
November or December in the sample.
Data Analysis
After the data were screened, time series line plots of the dependent variables
revealed periods of non-constant variance. However, a natural log transformation provided
a remedy for this condition. Figure 3 illustrates the critical line plots following the natural
log transformation. Any remaining peaks are the result of special events, holidays, and
phenomena reected on the right side of the regression equation. As endorsed by Kennedy
52
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
(1998, p. 264), these line plots were used to visually assess both rst and second order
stationarity (i.e., with regard to the series mean and variance). No unit root tests were
performed, as such tests have been repeatedly found to lack statistical power (Campbell
and Perron, 1991; Cochrane, 1991). However, each original model did include a linear
trend variable to account for any signicant changes in the mean of the series, over the
course of the sample period.
Time Series Plots: Resort 1
LN COIN Series (Top) & LN DROP Series (Bottom)
12
12.5
13
13.5
14
14.5
15
15.5
2/3/2009
2/17/2009
3/3/2009
3/17/2009
3/31/2009
4/14/2009
4/28/2009
5/12/2009
5/26/2009
6/9/2009
6/23/2009
7/7/2009
7/21/2009
8/4/2009
8/18/2009
9/1/2009
Time (in Days)
Natural Log of DVs
Time Series Plots: Resort 2
LN COIN Series (Top) & LN DROP Series (Bottom)
11
12
13
14
15
16
17
2/3/2009
2/17/2009
3/3/2009
3/17/2009
3/31/2009
4/14/2009
4/28/2009
5/12/2009
5/26/2009
6/9/2009
6/23/2009
7/7/2009
7/21/2009
8/4/2009
8/18/2009
9/1/2009
Time (in Days)
Natural Log of DVs
53
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Figure 3. Time series plots of dependent variables for Resorts 1, 2, and 3.
To ease interpretation of the nal model results, all continuous independent variable
values were also converted to their natural log form, creating double-log models. A
technique common in econometric modeling (Dielman, 1996), the double-log model
features regression coefcients which are expressed as elasticities, i.e., the elasticity of Y
with respect to X (Kahane, 2008, p. 84). Additionally, any replication of this double-log
model would express the poker room effect in a comparable metric (i.e., in the form of an
elasticity).
Following the natural log transformation of the continuous model variables,
descriptive statistics were reviewed along with relevant bivariate correlation coefcients.
The theoretical model illustrated in Figure 2 was tested via simultaneous-entry time series
regression analysis. All hypotheses were tested at the 0.05 alpha level. EViews, version
3.1, and SPSS, version 18 were used for all data screening, exploratory data analysis,
formal data analysis, and testing of methodological assumptions. The nal time series
models were produced in EViews, version 3.1.
Operalization of Dependent Variables
In total, there were six dependent variables, the natural log of aggregate daily coin-in
(COININ) and the natural log of aggregate daily table game drop (DROP), for each of the
three resorts. In the gaming industry, coin-in refers to the dollar amount of wagers placed
in some number of slot machines, over a given period of time. In this paper, “slots”
or “slot machines” refer to any coin- or voucher-operated device on the casino oor,
including but not limited to reel slots, video poker games, and electronic roulette games.
COININ did not include any wager placed in any gaming device that was permanently
attended to by a live dealer.
In the table game area of the casino (a.k.a. the pit), drop is the most widely used
business volume indicator. Although it is difcult to precisely describe what it actually
represents, drop remains the most common and usually best measure available for
describing the business volume in the pit. The following formula from Lucas & Kilby
(2012, p. 140) is the most effective way to describe/dene drop in a Nevada casino:
Time Series Plots: Resort 3
LN COIN Series (Top) & LN DROP Series (Bottom)
13
13.5
14
14.5
15
15.5
16
16.5
17
2/3/2009
2/17/2009
3/3/2009
3/17/2009
3/31/2009
4/14/2009
4/28/2009
5/12/2009
5/26/2009
6/9/2009
6/23/2009
7/7/2009
7/21/2009
8/4/2009
8/18/2009
9/1/2009
Time (in Days)
Natural Log of DVs
54
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Table Game Drop = Dollar-value of Currency in the Drop Box + Dollar-value
of Gaming Cheques in the Drop Box + Dollar-value of Marker Issue Slips in the
Drop Box – Dollar-value of Marker Redemption Slips in the Drop Box
This formula is used to compute the dollar-value of the drop box contents for each
table game. “Marker Issue Slips” refers to ticket-like forms detailing the issuance of
credit to a casino customer. “Marker Redemption Slips” represents the retirement of debt
issued by way of marker issue slips. Because marker redemption slips are included in
the formula, table game drop does not represent the dollar-amount of cheques purchased
at the tables (a.k.a. buy-in). This is a common mistake, when attempting to describe the
meaning of drop. There are other issues related to the use of buy-in as a proxy for drop,
but they are beyond the scope of this article.
Operalization of Independent Variables
The variable of interest in this study was RAKE, which represented the natural log
of the daily aggregate dollar amount of fees paid by poker players. Simply put, rake is
the fee charged to gamblers who play poker in the casino. Poker is somewhat unusual
in that the game has no house edge. With no casino advantage, management must levy
a rake against the players to make the game protable. Aside from hourly headcounts of
the poker room, which few management teams collect, rake represents the best available
business volume indicator for a poker room.
TREND is the natural log form of the variable designed to represent the presence
of any positive or negative linear trend in the dependent variable values, over the course
of the sample. This is equivalent to taking the natural log of a counter variable, which is
assigned a value of one on the rst day of the sample and increases by a value of one with
each additional day.
The day-of-the-week variables were expressed in a binary format. For example, on
Sunday, the Sunday variable was set to a value of one, with all other of day-of-the-week
variables set to a value of zero for that day. The nal day-of-the-week variables were
labeled as follows: TUE, WED, THU, FRI, SAT, and SUN. Monday was used as the base
period, providing a level from which all other day-of-the-week variables either did or did
not vary.
Like the day-of-the-week variables, the holiday and special event variables were also
expressed as binary variables. These variables were included to represent the days of a
particular holiday period or special event. The nal variables were labeled as follows:
INDDAY (Independence Day holiday), KDERBY (Kentucky Derby event), LABORDAY
(Labor Day holiday), MEMDAY (Memorial Day holiday), NCAABBALL (men’s college
basketball tournament event), PRESDAY (Presidents’ Day holiday), and STPATS (Saint
Patrick’s Day holiday).
Finally, appropriate ARMA terms were added to the models as needed, to produce
an independent error process. The “AR” stands for autoregressive and the “MA”
represents moving average. Common to time series models, these terms not only remove
serial correlation from the errors, they also represent an unnamed and often powerful
explanatory process within the error structure.
55
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Results
Descriptive Statistics
Once the data were screened, descriptive statistics were computed and examined.
These measures are listed in Table 1, stated in natural log form. Other than the key variable,
RAKE, Table 1 only includes the continuous variables appearing in the nal models for
each resort. Results for variables that failed to produce a statistically signicant effect are
not shown in any of the tables appearing in this section.
Table 1
Descriptive Statistics for Continuous Model Variables:
Resorts 1, 2, and 3 (n = 217)
Mean Std. Dev. Min. Max.
Resort 1 Variables:
COIN-IN 14.55 0.22 14.11 15.22
DROP 12.86 0.35 12.16 13.78
RAKE 8.68 0.31 7.59 9.38
Resort 2 Variables:
COIN-IN 15.67 0.36 14.99 16.65
DROP 14.76 0.54 13.67 16.43
RAKE 9.93 0.23 9.37 10.49
Resort 3 Variables:
COIN-IN 15.38 0.24 14.90 15.96
DROP 11.91 0.26 11.28 12.72
RAKE 8.26 0.25 7.64 8.95
TREND 4.40 0.96 0.00 5.38
Notes. All values are expressed as the natural log of the original metric.
From Table 1, the standard deviation of Resort 2’s DROP is noticeably greater than
its Resort 1 and 3 counterparts. This is most likely due to Resort 2’s high-roller table
game clientele, which is infamous for the volatility it creates. Table 2 lists the bivariate
correlation coefcients for the natural log of the continuous variables appearing in the nal
models.
Table 2
Correlation Matrices for Continuous Model Variables:
Resorts 1, 2, and 3 (n = 217)
COIN-IN DROP RAKE TREND
Resort 1 Variables:
COIN-IN --
DROP 0.79 --
RAKE 0.64 0.72 --
TREND n/a n/a n/a --
Resort 2 Variables:
COIN-IN --
DROP 0.57 --
RAKE 0.51 0.41 --
TREND n/a n/a n/a --
Resort 3 Variables:
COIN-IN --
DROP 0.83 --
RAKE 0.68 0.61 --
TREND -0.23 -0.22 -0.22 --
Notes. All correlation coefcients were signicant at the 0.01
alpha level (2-tailed). “n/a” represents not applicable,
as TREND was not present in the nal models for Resorts
1 and 2.
56
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Slot Models
Table 3 contains the results of the double-log models designed to predict COININ
at each of the three resorts. For Resort 1, the model generated an R
2
of 78.2% and an
F-statistic of 59.87 (df = 11, 205). The model was slightly more effective on the Resort
2 data set, posting an R
2
of 81.7% and an F-statistic of 59.35 (df = 14, 202). Finally, the
model was most successful in explaining the variation of COININ at Resort 3, producing
an R
2
of 89.9% and an F-statistic of 123.8 (df = 13, 203). As for the key variable, RAKE,
it failed to produce a signicant model effect in any of the three data sets (i.e., all three
p-values > 0.05).
Table 3
Results of Double-log Time Series Regression Analyses
Dependent Variable: Natural Log of Aggregate Daily Coin-in
Resort 1 Resort 2 Resort 3
Variable
B p B p B p
Constant 13.8990 14.5259 15.0887
RAKE 0.0646 0.0959 0.0991 0.4265 0.0570 0.1404
TUE -0.0368 0.0499 -0.0886 0.0017 n/a
WED -0.0626 0.0011 -0.1077 0.0084 0.1541 0.0000
THU n/a 0.1109 0.0112 0.1077 0.0000
FRI 0.2279 0.0000 0.4074 0.0000 0.5182 0.0000
SAT 0.3297 0.0000 0.5751 0.0000 0.4715 0.0000
SUN 0.1383 0.0000 0.3001 0.0000 0.1789 0.0000
TREND n/a n/a -0.0870 0.0002
PRESDAY n/a 0.6225 0.0000 0.2742 0.0000
STPATS 0.2650 0.0040 n/a 0.2637 0.0003
MEMDAY 0.2774 0.0006 n/a 0.1660 0.0044
INDDAY 0.2248 0.0053 n/a n/a
LABORDAY n/a n/a 0.1798 0.0166
FEB 22 n/a -0.2218 0.0443 n/a
MAR 12 0.2513 0.0061 n/a n/a
MAR 27 n/a -0.1567 0.0499
JUN 10 n/a n/a 0.3015 0.0001
JUN 19 0.2480 0.0070 n/a n/a
AUG 12 n/a 0.4267 0.0007 n/a
AR (1) 0.6257 0.0000 0.8797 0.0000 0.4610 0.0000
MA (2) n/a -0.2320 0.0129 n/a
MA (3) n/a -0.2274 0.0090 n/a
MA (7) n/a -0.1790 0.0098 0.2390 0.0001
R
2
78.23% 81.66% 89.89%
Model F-Statistic 59.87 0.0000 59.35 0.0000 123.82 0.0000
Notes. “n/a” represents not applicable, i.e., the variable did not appear in
the nal model. The VIFs for RAKE were 1.7, 2.1, and 2.1 for the Resort 1,
Resort 2, and Resort 3 models, respectively. A p-value of 0.0000 indicates
a value less than 0.00005. Per the Methodology section, all p-values < 0.05
indicate a statistically signicant effect, i.e., alpha = 0.05.
The following variables listed in Table 3 represented outlier dates: FEB 22, MAR 12,
MAR 27, JUN 10, JUN 19, and AUG 12. With a sample size of 217 observations from
each resort, the presence of outliers came as no surprise. Outliers were dened as dates
exhibiting a studentized deleted residual greater than 3.0. After carefully examining the
variable values for the outlier dates, there was no reason to doubt the legitimacy of the
observations. Therefore, binary indicator variables representing the unusual conditions on
these dates were created and included in the nal models.
Finally, the appropriate AR and MA terms were added to each of the models,
following a review of the correlograms associated with each model’s autocorrelation
function and partial autocorrelation function of the errors. Once the ARMA terms listed in
Table 3 were added to the models, the correlograms were examined a second time to
verify that the model errors were free from signicant serial correlation. This implied
that the independent variables and ARMA terms provided a good t to the data.
57
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Table Game Models
Table 4 includes the results of the double-log models designed to predict DROP at
each of the three resorts. The Resort 1 data t the model best, producing an R
2
of 91.0%
and an F-statistic of 184.15 (df = 10, 206). The Resort 2 model posted an R
2
of 65.5%
and an F-statistic of 29.54 (df = 12, 204), while Resort 3 generated an R
2
of 80.7% and
an F-statistic of 77.38 (df = 10, 206). Again, Resort 2’s relatively low R
2
value was most
likely a function of its high-roller table game clientele. On any given day, a single high-
roller can produce an extreme outcome capable of greatly affecting the casino’s aggregate
drop.
RAKE produced a signicant model effect at Resort 1 (B = 0.0886; p < 0.05).
That is, a 1% increase in RAKE produced an 8.86% increase in DROP. RAKE failed to
produce a statistically signicant effect in the Resort 2 and 3 models.
The identication, investigation, and treatment process of outliers was identical to
that described in the previous section. The ARMA terms were also specied according
to the protocol described in the previous section. However, each of the three table game
models required only the addition of an AR (1) term, otherwise known as an ARIMA
(1,0,0) model.
Table 4
Results of Double-log Time Series Regression Analyses
Dependent Variable: Natural Log of Aggregate Daily Table Game Drop
Resort 1 Resort 2 Resort 3
Variable
B p B p B p
Constant 11.8308 14.0277 11.8653
RAKE 0.0886 0.0268 0.0430 0.8565 0.0236 0.6599
TUE -0.0457 0.0058 n/a n/a
THU 0.1588 0.0000 0.2147 0.0015 0.0785 0.0008
FRI 0.5261 0.0000 0.6493 0.0000 0.4940 0.0000
SAT 0.7245 0.0000 0.7243 0.0000 0.4853 0.0000
SUN 0.3675 0.0000 0.4957 0.0000 0.1885 0.0000
TREND n/a n/a -0.0745 0.0002
STPATS n/a n/a 0.3032 0.0051
MEMDAY 0.2865 0.0007 n/a n/a
INDDAY n/a n/a n/a
LABORDAY n/a 0.7436 0.0096 n/a
NCAABBALL 0.1560 0.0449 n/a n/a
KDERBY n/a 0.5891 0.0234 n/a
FEB 4 n/a -0.5889 0.0394 n/a
FEB 11 n/a -0.5139 0.0496 n/a
MAR 10 n/a n/a 0.4990 0.0000
APR 8 n/a -0.7842 0.0058 n/a
APR 25 n/a n/a 0.4030 0.0002
MAY 15 n/a -0.9073 0.0017 n/a
JUN 1 n/a n/a 0.3701 0.0007
JUN 17 n/a 0.5689 0.0448 n/a
JUN 30 0.6393 0.0000 n/a n/a
JUL 13 -0.2605 0.0031 n/a n/a
AR (1) 0.7358 0.0000 0.6013 0.0000 0.5111 0.0000
R
2
90.97% 65.53% 80.67%
Model F-Statistic 184.15 0.0000 29.54 0.0000 77.38 0.0000
Notes. “n/a” represents not applicable, i.e., the variable did not appear in
the nal model. The VIFs for RAKE were 1.8, 2.1, and 2.1 for the Resort 1,
Resort 2, and Resort 3 models, respectively. A p-value of 0.0000 indicates
a value less than 0.00005. Per the Methodology section, all p-values < 0.05
indicate a statistically signicant effect, i.e., alpha = 0.05.
58
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Slot and Table Game Model Diagnostics
With regard to methodological assumptions, independence was addressed by
examining correlograms of the autocorrelation and partial autocorrelation functions
of the error process for each model. After the appropriate ARMA terms were added to
the models, the Q-statistics associated with the correlograms revealed no statistically
signicant serial correlation across 36 lags.
The linearity assumption was examined by reviewing scatter plots of the model errors
against each predictor variable series. The presence of nonconstant variance was also
assessed by way of the scatter plot. In this case studentized deleted residuals were plotted
against corresponding predicted values. With respect to both linearity and nonconstant
variance, the scatter plots revealed no cause for concern. Finally, the distributions of the
model errors were examined via histograms, which revealed no problematic departures
from the normal distribution.
Multicollinearity levels were assessed by reviewing the variance ination factors
(VIFs) for each set of predictor variables. As RAKE was the key variable in this study,
its VIFs were the primary concern. As listed in the notes of Tables 3 and 4, the VIFs for
RAKE ranged from 1.7 to 2.1 in the COININ models and from 1.8 to 2.1 in the DROP
models. These results indicated low to mild levels of multicollinearity. In fact, the greatest
VIF of any predictor variable in the COININ models was 2.5, with the DROP models
posting a high VIF of 2.1.
Discussion
Only one of six null hypotheses was rejected, indicating a considerable lack of
support for both the full service theory, in general, and the assumption of indirect revenue
generation by poker rooms, in particular. These results were consistent with the ndings
of previous researchers exploring the links between coin-in levels and other gaming
amenities such as race and sports books and bingo rooms (Abarbanel et al., 2011; Lucas
et al., 2006). However, these same results were not consistent with the ndings of Ollstein
(2006), with respect to the general effect of poker rooms on daily coin-in levels, and
Lucas and Brewer (2001), regarding the effect of bingo headcount on daily coin-in.
Including this work, there have been ve studies that have empirically examined the
relationships depicted in the Casino block of Figure 1. In total, these studies have tested
13 Casino block hypotheses, using secondary data from eight hotel-casinos. The results
of ten of the 13 hypothesis tests have failed to support the full service theory within the
Casino block. Operators may want to consider this battery of results, when discussing and/
or estimating the indirect revenue contributions of gaming amenities such as poker rooms,
race and sports books, and bingo rooms. This is not to say that further research is not
needed, but the collective results are considerably one-sided.
The ndings of the current study also supported the notion of cherry picking, in
that poker players seemed to lack interest in slots and table games. Specically, the lack
of signicant effects for RAKE suggested a general unwillingness to patronize more
protable areas of the casino, especially the most protable area – slots. Staying with
the retail literature, these same results failed to support the spillover effect. That is, the
ndings did not cast the poker room as a draw that produces business for other key areas
of the casino.
Finally, the outcome of this work was at odds with the popular opinions of industry
insiders regarding the ability of poker rooms to drive business to slots and table games
59
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
(Cosgrove-Mather, 2005; Grochowski, 2005; Legato, 2010; McGowan, 2010; Taucer,
2004; Walters, 2003; Wiser, 2004). As academic research continues to produce conicting
empirical results, the trade journal consensus on the full service theory grows increasingly
mysterious. At some point, these insiders must show their cards. Exactly what is it that
suggests a signicant and positive relationship between gaming amenities and key gaming
business volumes?
From the trade literature, it would seem as though many operators are content in
assuming the full service theory is valid until proven otherwise. This may be a dangerous
and costly default position, given the paucity of objectively and rigorously derived
empirical support for the full service theory. While it is very difcult to identify the origin
of the full service theory, it seems to survive on the legs of simplistic cross-tabulations,
casual observations, fear of extinction, and a general resistance to change.
Managerial Implications
The ndings are particularly troubling for Resorts 2 and 3. If these poker rooms were
producing insufcient prots per square foot, then executives would be wise to entertain
alternative uses of the oor space. For these operators, continued justication of live
poker by way of indirect revenue contributions may have just become a tougher sell. Of
course, additional research is recommended before making any decisions. Alternative
approaches are described in the upcoming Future Research section.
As for Resort 1, the bulk of its operating prots came from slots and the hotel. Both
of these prot centers typically feature impressive prot margins (Lucas & Kilby, 2012).
Although RAKE posted a signicant and positive effect in Resort 1’s table game drop
model, the regression coefcient reected the increase in drop. Of course, drop must be
converted into win and ultimately into operating prots. This conversion process is subject
to the inated variable cost structure of the Table Games Department, eroding much of the
contribution reected in the regression coefcient (i.e., the 8.86% increase).
At a minimum, those faced with poker room addition, deletion, or modication
decisions may consider decreasing the weight of any indirect revenue gains/losses
included in their projections. This recommendation holds for operators and developers
alike. The demonstrated lack of empirical support for the full service theory in general,
and as it applies to poker rooms, must be considered in these important decisions.
Limitations
The results of this work cannot be generalized beyond the casino oors of the donor
properties. Further, including Ollstein (2006), poker data from only four different hotel-
casino resorts have been examined. All three of the samples examined in this research
began in early February and extended through early September. Ollstein featured a nearly
identical sample period. Although there is no reason to believe so, it is possible that the
indirect revenue contributions of poker rooms are different in the period ranging from
mid-September to February.
The econometric models used in this research are often referred to as causal models.
However, time series regression analysis does not prove cause and effect. Its use herein
only served to test the plausibility of the theoretical model advanced. Proving cause and
effect is not a statistical question.
60
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Future Research
While replication of the current model would supply valuable results to the research
stream, alternative measures of poker room business volume would provide a more
diverse representation of possible indirect revenue contributions. One such measure
would be an hourly headcount of poker players, to be used in lieu of daily rake. Hourly
headcounts analyzed in conjunction with hourly coin-in would allow for a more
temporally congruent measurement of the relationship between the business volumes of
the poker room and the slot oor. The results of the current study may provide the push
that is needed to gain support for the study of hourly business volumes.
A before-and-after look at a property featuring a poker room closure would provide
an alternative yet useful perspective. For example, a times series model could examine
daily business volume data for several months before and after the closure. The poker
room variable could be expressed in a binary format, producing a regression coefcient
that would represent the offset/value of having the poker room, expressed in terms of
key gaming volumes such as coin-in and table game drop. There are several Las Vegas
properties with data to accommodate this approach, given the recent closure of several
poker rooms.
An observational study could be conducted to better understand the crossover effect
as it applies to poker players. For example, how many poker players relocate at a gaming
position after leaving the poker room? How long do they play slots and/or table games?
How much do they wager? Such a study would also contribute to a more complete
understanding of the poker player’s total value to the casino. However, it would be much
more difcult to conduct an observational study of poker players prior to their arrival in
the poker room. Observers would have to be able to identify poker players as they enter
the casino or be willing to follow a great number of subjects to locations other than the
poker room.
Finally, the relationship between the poker room and protable nongaming amenities
could be examined. The results of the current study will surely produce claims of
contributions to nongaming amenities. Any research aimed at these claims should focus
on estimating the relationship between the poker room and the hotel or something along
the lines of a hyper-protable nightclub such as the Cosmopolitan’s Marquee or Encore’s
XS. When it comes to operating prots, not all nongaming amenities are alike. For
example, few if any come close to matching the prots of the hotel operation (Lucas and
Kilby, 2012). To the contrary, demonstrating a positive relationship with restaurant or
showroom business volumes is not likely to save the poker room, as neither of these areas
would be considered a prot juggernaut.
References
Abarbanel, B. L. L., Lucas, A. F., & Singh, A. K. (2011). Estimating the indirect effect of
sports books on other in-house gaming volumes. UNLV Gaming Research & Review
Journal, 15(2), 77-90.
Arnett, K.. (2010). The Las Vegas grinder: Update on the Hard Rock and the lowdown on
room rates. Retrieved on February 29, 2012, from http://www.pokernews.com/
news/2010/ 08/the-vegas-grinder-update-on-the-hard-rock-and-the-lowdown-
on-8810.htm
61
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Boncek, C. (2008). Hard Rock poker lounge open for business. Retrieved on May 1,
2012, from http://www.pokernews.com/news/2008/08/hard-rock-poker-lounge-open.
htm
Brock, F. J., Fussell, G. L., & Corney, W. J. (1990). How to optimize casino-hotel
revenue. Journal of Business Forecasting, 9(2), 2-5, 10.
Byrne, T. (2010). Early results on state’s new table games: Casinos bet revenues will
climb. Retrieved on November 26, 2010, from http://www.delawarerst.org/
government_and_politics/early-results-on-states-new-table-games-casinos-bet-
revenues-will-climb/
Campbell, J. Y., & Perron, P. (1991, March 8-9). Pitfalls and Opportunities: What
Macroeconomists Should Know about Unit Roots. Proceedings of the NBER
Macroeconomics Conference, Cambridge, MA.
Cochrane, J. H. (1991). A critique of the application of unit root tests. Journal of
Economic Dynamics and Control, 15, 275-284.
Cosgrove-Mather, B. (2005). Despite limited prots, casinos repackaging poker as
attraction. Retrieved on November 22, 2010, from http://www.cbsnews.com/
stories/2005/04/11/entertainment/main687321.shtml
Dandurand, L., & Ralenkotter, R. (1985). An investigation of entertainment proneness and
its relationship to gambling behavior: The Las Vegas experience. Journal of Travel
Research, 23(3), 12-16.
Dielman, T. E. (1996). Applied Regression Analysis for Business and Economics. (2nd
ed.). Duxbury Press, New York.
Eng, R. (2008). Vegas race books buck trend: Millions of dollars in upgrades prompt
increase in pari-mutual wagering. Las Vegas Review Journal. Retrieved November
14, 2009 from http://www.lvrj.com/sports/14309912.html
Eppli, M. J., & Shilling, J. D. (1995). Large-scale shopping center development
opportunities. Land Economics, 71(1), 35-41.
Geller, R. (2009). How technology is improving the prots for poker rooms. Global
Gaming Business, 8(6), 28, 30-31.
Grochowski, J. (2005). Poker rooms ride high with TV boom. Casino
Answer Man. Retrieved February 27, 2010, from http://www.casinoanswer
man.com/ articles_tvboom.html
Hanks, R. B., Noland, R. P., & Cross, R. G. (1992). Discounting in the hotel industry: A
new approach. Cornell Hotel & Restaurant Administration Quarterly, 33(3), 40-45.
Kahane, L. H. (2008). Regression Basics. (2nd ed.). Sage, Thousand Oaks, CA.
Kennedy, P. (1998). A Guide to Econometrics. (4th ed.). MIT Press, Cambridge.
Kimes, S. E., & Thompson, G.M. (2004). Restaurant revenue management at Chevys:
Determining the best table mix. Decision Sciences, 35(3), 371-391.
62
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Legato, F. (2010). Back to the pit? Global Gaming Business, 9(8), 28-32.
Lucas, A. F. (2004). Estimating the impact of match-play promotional offers on the
blackjack business volume of a Las Vegas hotel casino. Journal of Travel &
Tourism Marketing, 17(4), 23-33.
Lucas, A. F. (2010). Estimating untracked gaming volumes by hotel occupancy segment.
Cornell Hospitality Quarterly, 52(2), 209-218.
Lucas, A. F., & Brewer, K. P. (2001). Managing the slot operations of a hotel casino
in the Las Vegas locals’ market. Journal of Hospitality & Tourism Research,
25(3), 289-301.
Lucas, A. F., Dunn, W. T., & Kharitonova, A. (2006). Estimating the indirect gaming
contribution of bingo rooms. Gaming Research & Review Journal, 10(2), 39-54.
Lucas, A. F., & Kilby, J. (2008). Principles of Casino Marketing. San Diego, CA: Gamma.
Lucas, A. F., & Kilby, J. (2012). Introduction to Casino Management. San Diego, CA:
Gamma.
Lucas, A. F., & Santos, J. (2003). Measuring the effect of casino-operated restaurant
volume on slot machine business volume: An exploratory study. Journal of
Hospitality & Tourism Research, 27(1), 101-117.
Manteris, A. (1993). What went wrong at Sport of Kings: The Sports Book is primarily to
bring business to the casino. Gaming & Wagering Business, 63.
McGowan, R. A. (2010). The poker boom: Can casinos cash-in on low-stakes Texas
Hold’Em? Gaming Law Review and Economics, 14(7), 525-531.
Mehaffey, J. (2012). Consolidation of Las Vegas poker rooms. Retrieved on February 2,
2013, from http://www.legalpokersites.com/blog/consolidation-of-las-vegas-poker-
rooms/2929/
Ollstein, B. (2006). Estimating the indirect gaming contribution of poker rooms: Is poker
room volume a peripheral driver of slot revenue? [Unpublished Masters Thesis].
University of Nevada, Las Vegas.
Roehl, W. S. (1996). Competition, casino spending, and use of casino amenities. Journal
of Travel Research. 34(3), 57-62.
Sieroty, C. (2011, March 8). Gold Coast moves poker room to the Orleans. Las Vegas
Review Journal. Retrieved on March 4, 2012, from http://www.lvrj.com/
business/gold-coast-moves-poker-room-to-the-orleans-117610748.html
Smith, B. A., Leimkuhler, J. F., & Darrow, R. M. (1992). Yield management at American
Airlines. Interfaces, 22(1), 8-31.
Suh, E. (2006). Estimating the impact of entertainment on the gaming volume of Las
Vegas hotel casinos. [Unpublished Doctoral Dissertation]. University of Nevada, Las
Vegas.
Taucer, V. (2004). Table games experience growth in all quarters. International Gaming
& Wagering Business, 74(October).
63
UNLV Gaming Research & Review Journal w Volume 17 Issue 1
Taucer, V. (2005). Bet carefully with poker. International Gaming & Wagering Business,
32(April).
Walters, R. (2003). Know when to hold ‘em. Global Gaming Business, 2(16), 40-43.
Walters, R. G., & MacKenzie, S. B. (1988). A structural equations analysis of the impact
of price promotion. Journal of Marketing, 25(1), 51-63.
Walters, R. G., & Rinne, H. J. (1986). An empirical investigation into the impact of price
promotions on retail store performance. Journal of Retailing, 62(3), 237 –266.
Wiser, R. (2004). Ante up! Casino Player, 15(9), 85-88, 90.
Wolf, N. (2010). Beyond brick and mortar: A whole new poker room (.com). Global
Gaming Business, 9(11), 24-26.
