The Market for Online Poker
Ingo Fiedler
Ann-Christin Wilcke
Abstract
The recent events of the “Black Friday” – the biggest online poker networks in
the USA were shut down – demonstrated the necessity to make decisions about the
regulation of online poker. But although online poker is a gold mine of data, until
now nobody knows where the players and their money come from. It seems that the
knowledge about the online poker market has not been able to keep up with the speed
of its evolution in the past years. This paper is the first to shed light on this matter. We
use data of 4,591,298 poker identities from the Online Poker Database of the University
of Hamburg (OPD-UHH) collected over a six months period from 09/2009 to 03/2010.
We find that the worldwide 6 million players paid 3.61 billion US$ rake to the operators
in 2010. USA is still by far the biggest market with 1,429,943 active players and 973.3
million US$ net revenues in 2010. With regard to the number of internet users in a
country, Hungary is the biggest relative market: One out of 50 Hungarians with internet
access plays online poker for real money. The two main drivers of the relative market size
in a country are GDP per capita and culture. Using the data from the OPD-UHH, future
research will be able to break down the market also on a regional level within countries
and to examine inter and intra country differences in the playing habits of online poker
players.
Keywords: online, poker, market, gambling, data, habits.
Ingo Fiedler
Research Associate
Institute of Law &
Economics
University of Hamburg
Ann-Christin Wilcke
Research Associate
Institute of Law &
Economics
University of Hamb
Introduction
Online gambling and online poker in particular is a relatively new phenomenon
nearly nonexistent in 2003. During the past years it has grown extraordinarily so that it is
now an important factor in the whole gambling market. However, even today it is often
neglected by the old industry, the legislators and the researchers as an unwelcome black
market just as if denying would cease its existence. But the economic reality is different.
Online gambling has tremendous cost advantages due to electronic instead of brick and
mortar operation and, even more important, its operators do not have to pay (high) taxes
and license fees. In the case of poker, large player pools and the corresponding network
effects have also helped the game to grow to a size not to be matched in the offline world.
From an academic perspective, most research in the field of poker has focused on
playing strategy (e.g. Chen & Ankenman, 2006) and especially the question whether
poker is a game of skill or a game of chance (see e.g. Dreef et al. 2003, Cabot & Hannum
2005, Dedonno & Detterman 2008, or Turner 2008 and for an empirical point of view see
Fiedler & Rock 2009). The answer to this question is especially important from a legal
perspective as most jurisdictions legalize games of skill while they regulate games of
chance. However, the skill/chance debate has not yet come to an end, and also skill does
not seem to be the best criterion for the distinction between harmful and non-harmful
games (Rock & Fiedler 2008).
The recent events of the “Black Friday” – the biggest online poker networks in
the USA were shut down – demonstrated the necessity to make decisions about the
regulation of online poker. Should it be prohibited? If so, what are the best instruments
to enforce a prohibition? Or should online poker be legalized like Italy and France
have chosen lately? If so, what is the optimal tax and how could player protection be
implemented? Before answering these questions it may be worthwhile to gather more
information about the national markets as information about them is practically nonUNLV Gaming Research & Review Journal ♦ Volume 16 Issue 1
7
existent.
We use data from the Online Poker Database of the University of Hamburg (OPDUHH), a database including information on the origin and playing habits of 4,591,298
poker identities over a six months period. Before describing the
data in the OPD-UHH we analyze the worldwide online poker
The recent events of the “Black
market and the market shares of the various poker operators. We
then present the results about the size of the whole market and for
Friday” – the biggest online
the ten biggest markets in dollar terms. Afterwards we analyze
poker networks in the USA were
the prevalence of online poker in absolute terms in relation to the
shut down – demonstrated the
internet users. We also identify the main drivers of the relative
market size of a country. Both sections provide new and important necessity to make decisions about
insights into the online gambling market especially for legislators.
the regulation of online poker.
We continue by highlighting the limitations of this study before we
summarize the results and give an outlook on further research in the
last section.
Market Size and Market Shares by Operators
With the exception of an analysis of poker marke ts in the European Union (Fiedler
& Wilcke 2012) academic research on the prevalence and size of online poker markets
thestill
observation
of the
industry by
PokerScout
which tacks the
players atby
thePokerScout
different operators
at
is
missing.
However,
due
to the observation
of active
the industry
which
tacks
themoment
active the
players
the different
operators
at any
given
moment
the market
any given
marketatshares
of the operators
are known.
Table
1 shows
the market
shares forshares
the
of
operators
are inknown.
Table
shows
market
theare
nine
biggest
ninethe
biggest
poker sites
2008, 2009
and12010
(untilthe
August).
Theshares
marketfor
shares
computed
in poker
sites
in
2008,
2009
and
2010
(until
August).
The
market
shares
are
computed
in
relation
relation to the monthly average players of the total market: 53,121 in 2008, 68,483 in 2009 and 71,441 in
to the monthly average players of the total market: 53,121 in 2008, 68,483 in 2009 and
2010. Note that a player is not counted in this statistic unless he is playing at the moment of the scan by
71,441 in 2010. Note that a player is not counted in this statistic unless he is playing at
PokerScout. The average number accounts for all differences in active players per daytime, week days
the
moment of the scan by PokerScout. The average number accounts for all differences
andactive
months.players per daytime, week days and months.
in
Table 1
Market shares in the online poker industry in 2008, 2009, and 2010
Site/Network
Pokerstars
Full Tilt Poker
iPoker Network
Party Poker
Cereus Networkb
Everest Poker
Microgamingc
IPN (Boss Media)d
Cake Poker Network
Ongame (bwin)e
Other
2008
2009
2010a
Market share
30.53%
14.65%
9.82%
8.72%
2.35%
4.85%
1.84%
2.81%
2.18%
6.86%
15.39%
Market share
36.05%
19.45%
8.40%
7.22%
3.40%
3.36%
2.71%
2.80%
2.45%
3.96%
10.20%
Market share
40.96%
21.73%
5.77%
6.12%
3.03%
1.80%
2.66%
2.88%
1.97%
3.55%
9.53%
Data recorded for
OPD-UHH
Yes
Yes
No
No
No
Yes
No
Yes
Yes
No
No
a
: 2010 includes only data from January to August.
: Cereus merged with Ultimate Bet in 12/2008.
c
: Microgaming merged with Ladbrokes in 03/2009.
d
: IPN (Boss Media) merged with Cryptologic 04/2009.
e
: Ongame (bwin) merged with Betfair in 08/2010.
b
Theonline
online
poker
industry
is dominated
highly dominated
by theproviders:
two biggest
providers:
The
poker
industry
is highly
by the two biggest
Pokerstars
and Full Tilt
Pokerstars
and
Full
Tilt
Poker,
and
their
importance
has
increased
over
the past years.
Poker, and their importance has increased over the past years. In 2008 they had a combined market share
In 2008 they had a combined market share of 45.18% while in 2010 they already
of 45.18% while in 2010 they already accounted for nearly two thirds (62.69%) of the whole playing
accounted for nearly two thirds (62.69%) of the whole playing volume in the market.
volume in the market. the
Correspondingly,
the smaller
lost significance
and dwindled
not only
in
Correspondingly,
smaller operators
lostoperators
significance
and dwindled
not only
in relative
relative size but also in absolute size. While in 2008 on the average 29,121 players were actively playing
8on smaller UNLV
Gaming
Research
& Review
Journal
♦ Volume
16 Issue
1
networks
at any given
moment,
the number
decreased
to 26,655
in 2010.
3 size but also in absolute size. While in 2008 on the average 29,121 players were actively
playing on smaller networks at any given moment, the number decreased to 26,655 in
2010.
This development shows the power of positive network effects in the online poker
market. The bigger the player pool, the more players are attracted to it just by its mere
size. As a player, you will more likely find (weak) opponents on a large network for
your favorite game type, betting structure and limit. This is even more true for players
who play multiple tables at the same time (“multitabling”) and therefore generate
significantly more traffic than players who only play at one table. Another important
factor reason for the large and increasing market share of PokerStars is the price of the
product in form of the rake structure, which is cheapest for the players at Pokerstars. But
this only applies to the gross costs without considering rakeback or promotions which
are used to set incentives for players to play more often and stick to one site. Rakeback
andThis
bonuses
effectively reduce the costs for the players and are, hence, also an important
development shows the power of positive network effects in the online poker market. The bigger
factor which drives market shares. Other factors include the quality of the software,
the player pool, the more players are attracted to it just by its mere size. As a player, you will more likely
security, reputation, ease of deposits and withdrawals, and also interactions between
find online
(weak) opponents
on a large
network(e.g.
for your
favoriteinto
game
type,
betting
structure
and limit.
Thisofis
the
site and offline
casinos
satellites
live
events
like
the Word
Series
even more true for players who play multiple tables at the same time (“multitabling”) and therefore
Poker).
Knowing
the market
shares
ofplayers
the different
next
step important
is to lookfactor
at the
generate
significantly
more traffic
than
who onlyoperators,
play at one the
table.
Another
revenue
which
is
generated
in
the
market.
The
gambling
consultant
company
H2GC
reason for the large and increasing market share of PokerStars is the price of the product in form of the
estimated the size of the online poker market in 2009 to be 2.4 billion US$ (H2GC
rake structure, which is cheapest for the players at Pokerstars. But this only applies to the gross costs
2009). Another way to determine the revenues is to look at the revenues stated in
without considering
rakeback
promotions
which and
are used
to set incentives
for players
more
financial
reports of
stock or
listed
companies
extrapolate
it in relation
to to
itsplay
market
often and
stick
to one site.
Rakeback andreported
bonuses effectively
reduce the
costs forwith
the players
and are,
share.
For
example,
PartyGaming
to have earned
together
its subsidiary
PartyPoker,
196.7
million
US$
net
revenues
in
2009
(Partygaming
2010).
Extrapolated
hence, also an important factor which drives market shares. Other factors include the quality of the
with
the 7.22% average market share of PartyPoker in 2009, the total market for online
software, security, reputation, ease of deposits and withdrawals, and also interactions between the online
poker was 2.7 billion US$, close to the estimate of H2GC.
site and offline casinos (e.g. satellites into live events like the Word Series of Poker).
But these figures are only estimates. They also do not answer where the money
comes
from and which the biggest national markets with the most players are. What
Knowing the market shares of the different operators, the next step is to look at the revenue which is
was the size of the US-market before the Black Friday? Was it still the largest market
generated in the market. The gambling consultant company H2GC estimated the size of the online poker
although the legislator tried to enforce a prohibition of online gambling with the
market in 2009
to be 2.4
billion US$
(H2GC 2009).Act
Another
way tosince
determine
the revenues
is to
look at
Unlawful
Internet
Gambling
Enforcement
(UIGEA)
2006?
These are
questions
the revenues
reports
stock listed companies
anddescribe
extrapolate
it in relation to its
we
address stated
with in
thefinancial
data from
theofOPD-UHH
which we
next.
market share. For example, PartyGaming reported to have earned together with its subsidiary PartyPoker,
The
Online Poker Database of the University of Hamburg (OPD-UHH)
196.7 million US$ net revenues in 2009 (Partygaming 2010).1 Extrapolated with the 7.22% average
Online poker is a goldmine of data as all operators show in their lobby a lot of
market share of PartyPoker in 2009, the total market for online poker was 2.7 billion US$, close to the
information about the people playing at their tables. It is easy to observe the origin
2
estimate
of H2GC.
of
a player
(city or country ), the game type, betting structure and limit of the table
they play at, and, of course, the date and time. Financed by the city of Hamburg, the
But these figures are only estimates. They also do not answer where the money comes from and which
Institute
of Law & Economics at the University of Hamburg collected these data in
the biggest national
markets
with the mostmarket
players are.
What was
the size of the
before the
collaboration
with
the independent
spectator
PokerScout
in US-market
the OPD-UHH.
A
software
electronically
gathered
the data
of the
thetofollowing
poker networks:
Black Friday?
Was it still the
largest market
although
theplayers
legislatorfor
tried
enforce a prohibition
of online
Pokerstars,
Full
Poker,
Everest
Poker,Enforcement
IPN (BossAct
Media)
andsince
Cake
Poker.
This
the Tilt
Unlawful
Internet
Gambling
(UIGEA)
2006?
These
are
gambling3 with
software
scanned
each
cash
game
table
of
the
mentioned
poker
sites
and
wrote
the
questions we address with the data from the OPD-UHH which we describe next.
displayed information into a SQL database.
The data collection was conducted for each poker site during a period of six months.
It took about ten minutes to scan all tables of an operator and collect the information
about the players sitting at the tables. This implies about 6 data points per hour or 25,920
in the course of six months and allows not only to determine the session lengths of the
players but also to analyze differences in time. The main period of the data collection
1
The gross revenue was 250.9 million US$. 54.2 million US$ were paid in bonuses to the players (so called
“rakeback”), meaning that the players got an average effective discount on the price of 21.6%.
2
Note that the extrapolation assumes that the revenues per player of PartyPoker are representative for the industry.
3
As the UIGEA applies to the definition of remote gambling in the Wire Act it does not necessarily include poker.
UNLV Gaming Research & Review Journal ♦ Volume 16 Issue 1
94 The data collection was conducted for each poker site during a period of six months. It took about ten
minutes to scan all tables of an operator and collect the information about the players sitting at the tables.
This implies about 6 data points per hour or 25,920 in the course of six months and allows not only to
determine the session lengths of the players but also to analyze differences in time. The main period of
the data collection took place during September 2009 and March 2010. Table 2 depicts the exact starting
took place during September 2009 and March 2010. Table 2 depicts the exact starting and
and ending
points
each site.
site. Note
period
of data
collection
had to be
extended
due to technical
ending
points
forforeach
Notethat
thatthethe
period
of data
collection
had
to be extended
due
problems
suchproblems
as downs of
the server,
software
updates
andsoftware
disconnections.
to
technical
such
as downs
of the
server,
updates and disconnections.
Table 2
Starting and ending points of the data collection for each site
Poker site
PokerStars
Full Tilt Poker
Everest Poker
IPN (Boss Media)
Cake Poker
Start
09/10/2009
09/06/2009
08/13/2009
07/27/2009
11/01/2009
End
03/11/2010
03/11/2010
03/11/2010
02/02/2010
07/02/2010
The Online Poker Database of the University of Hamburg (OPD-UHH)
In
the poker
6 months
of the data
collection
we obtained
data
for a4,591,298
pokerabout
identities
Online
is a goldmine
of data
as all operators
show in their
lobby
lot of information
the
including
country
of
origin
and
playing
habits.
Regarding
the
market
shares
of
In the 6 their
months
of the data
collection
wetheir
obtained
data for
4,591,298
poker identities
including
their
4
people playing at their tables. It is easy to observe the origin of a player (city or country ), the game type,
the
observed
poker
sites
these
are
64.72%
of
all
poker
players
worldwide
(88.27%
of
the
country of origin and their playing habits. Regarding the market shares of the observed poker sites these
betting structure and limit of the table they play at, and, of course, the date and time. Financed by the city
US-players and 57.4% of the players from all other countries) who were playing during
of Hamburg,
thethe
Institute
Law & Economics
at the University
of Hamburg
data inwe
the
period of
data of
collection.
If this number
is extrapolated
tocollected
the totalthese
market,
collaboration
with
independent
market spectator
PokerScout Note
in the OPD-UHH.
A software
get
a number
ofthe
7,094,095
different
player identities.
that one real
person can have
4
multiple
player
identities
if
he
opens
accounts
with
more
than
one
operator
many
electronically
gathered
data Everest
of the players
for IPN
the following
poker
networks:
Pokerstars,
Full
Tilt
The poker sites
Full Tiltthe
Poker,
Poker and
(Boss Media)
quote
the country
of originand
while
Poker Stars
and Cake have
Poker tried
specifyout
the city
of origin
of
the poker
player. site and therefore have multiple accounts.5
players
more
than
one
Poker,
Everest Poker, IPN (Boss Media) and Cake Poker. This software scanned each cash game table of
5
Play money tables have not been taken into consideration.
However,
to be counted
multiple
times
in theapproach
OPD-UHH
he collection
has to have
played with these
6
6
For
a more detailed
regarding
the technical
the adata
see
Sakai/Haruyoshi 2005.
the
mentioned
pokerdescription
sites and wrote
the displayed
informationofinto
SQL database.
accounts actively during the data collection period. This may be true for three groups of
players:
1) High volume players (professional and addicted gamblers), 2) Players on the 5 The data collection was conducted for each poker site during a period of six months. It took about ten
high
limits as they may need more than one site to find opponents and 3) People trying
minutes to scan all tables of an operator and collect the information about the players sitting at the tables.
out different sites to see which one suits them best. The high volume players have a huge
This impliestoabout
data
points
hour or 25,920
the course of six months
andget
allows
not onlywith
to
incentive
play6at
just
oneper
network
as the in
bonuses/rebates
they can
increase
determine
the session
lengths
of the on
players
also toare
analyze
in time.
Thereason
main period
their
playing
volume.
Players
highbutlimits
quitedifferences
rare (which
is the
whyofthey
may
have
to switch
to findSeptember
enough competition).
the data
collection
took sites
place during
2009 and March The
2010.data
Tablereveals
2 depictsthat
the for
exactexample
starting
only
0.36%
of
the
No
Limit
Holdem
players
are
on
high
stakes
(see
table
11
in
the
and ending points for each site. Note that the period of data collection had to be extended due to technical
appendix).
We therefore reason that most people with multiple player identities in our
problems such as downs of the server, software updates and disconnections.
sample come from the third group. We estimate that for 100 player identities about 85
real
players exist. This results in 6,029,980 people with a 6-months prevalence rate of
Table 2
participating in online poker cash games for real money.
Starting and ending points of the data collection for each site
Results: Market Sizes, Prevalence, and Drivers of Online Poker
The data in the OPD-UHH allows us to specify how many players in each country in
Poker
site play online poker for real moneyStart
the
world
and how much rake they paidEnd
to the operators
PokerStars
09/10/2009
03/11/2010
while playing. Using these data we determined the market sizes of the
different countries
Fulldollar
Tilt Poker
in
terms and also the absolute and09/06/2009
relative prevalence of online03/11/2010
poker in a country.
Everest Poker
IPN (Boss Media)
Market Sizes
Cake Poker
08/13/2009
07/27/2009
11/01/2009
03/11/2010
02/02/2010
07/02/2010
The OPD-UHH covers how often and how long a poker player played and if he
played multiple tables at the same time. We linked this information to data about the
average
rakeofpaid
percollection
player, hour,
and table
for4,591,298
the different
poker variants
In the US$
6 months
the data
we obtained
data for
poker identities
includingand
their
stakes.
This
allowed
us
to
determine
how
much
US$
rake
a
player
paid
to
the operator
country of origin and their playing habits. Regarding the market shares of the observed poker sites these
over the course of 6 months. As we also obtained the origin of the players, we could
aggregate the rake paid for all observed players of each country. We then extrapolated
4
The poker sites Full Tilt Poker, Everest Poker and IPN (Boss Media) quote the country of origin while Poker Stars
and Cake Poker specify the city of origin of the player.
Play money tables have not been taken into consideration.
6
For a more detailed description regarding the technical approach of the data collection see Sakai/Haruyoshi 2005.
5
10
UNLV Gaming Research & Review Journal ♦ Volume 16 Issue 1
5 this data to the whole market of each country by considering the market share of the
observed operators in each country (88.4% in the USA, 57.4% in all other countries,
resulting in 64.72% of the whole market). Multiplied by 2 we got the market size
for online poker cash games for each country in 2010. Given that the operators earn
about 70% of their revenues with cash games and 30% with tournaments we further
extrapolated the market size to include the tournament revenues. The resulting gross
market size is the rake paid by all online poker players in a given country. This is not
identical with the players’ net losses: 1) Players get about 25-30% of
Despite the prohibition of online their rake back in form of rakeback deals or bonuses and 2) players
can win and lose money from other players, a cash flow we did not
gambling and the introduction observe. To obtain the net losses, it is only necessary to account for
of the Unlawful Internet the former point, which can be easily done be deducting 30% of the
Gambling Enforcement Act gross market size. The latter point does not matter much for whole
countries, because cash flows between the players converge to zero if
(UIGEA) in 2006 the
US was
The OPD-UHH
covers
how
often andplayer
how long
a poker
player (without
played andrake,
if he played
tables
the
observed
pool
are large
pokermultiple
is a zero
sum
at the samemarket
time. We linked
this information to data about the average US$ rake paid per player, hour,
game).
still by far the biggest
10
Table
3 shows
the ten
biggest
poker
themuch
world
dollar
This
allowed
us to markets
determinein
how
US$inrake
a
and tablein
for2010.
the different poker
variants
and stakes.
with 973.3 million US$
terms
of
their
gross
market
size
and
their
share
of
the
total
market.
player paid to the operator over the course of 6 months. As we also obtained the origin of the players, we
Despite the prohibition of online gambling and the introduction of
could aggregate the rake paid for all observed players of each country. We then extrapolated this data to
the Unlawful Internet Gambling Enforcement Act (UIGEA) in 2006 the US was still by
the whole market of each country by considering the market share of the observed operators in each
far the biggest market with 973.3 million US$ in 2010. After the Black Friday, when the
country
(88.4%
in the USA, 57.4%
in all
other Absolute
countries, resulting
in 64.72%
of theBet
whole
market).
US sites
of Pokerstars,
Full Tilt
Poker
Poker and
Ultimate
were
shut down,
11
the US market
hasmarket
dropped
dramatically.
The games
remaining
operators
Cake
Poker
Given
that
Multiplied
by 2 wesize
got the
size for
online poker cash
for each
country inlike
2010.
12 10% before
or
Bodog
Poker
which
still
accept
US-players
had
a
market
share
of
less
that
the operators earn about 70% of their revenues with cash games and 30% with tournaments we further
the Black Friday. It will be interesting to see if these operators now rise and which level
extrapolated the market size to include the tournament revenues.13 The resulting gross market size is the
the US market will settle on.
rake paid by all online poker players in a given country. This is not identical with the players’ net losses:
The second largest market is Germany with a market size of 392 million US$. Other
14
and
2) players
can
1)
Players
get about
25-30% of
their rake
back in and
formGreat
of rakeback
dealsAlthough
or bonusesthe
big
markets
are France,
Russia,
Canada
Britain.
French
market
win
and
loselook
money
from other
players,
a cash
flow was
we did
not observe.
To obtain
the net
losses,
it is only
may
now
different
after
online
poker
legalized,
license,
taxed,
and
insulated
online
poker
was
legalized,
license,
taxed,
and
insulated
from
the
other
markets
in
2010.
In
total,
the
from
the
other
markets
in
2010.
In
total,
the
players
paid
3.6
billion
US$
rake
to
necessary to account for the former point, which can be easily done be deducting 30% of the grossthe
market
players paid 3.6
billion
US$
rake to
the operators
in 2010 or 599 US$ per player.
operators
in point
2010does
or 599
US$
per player.
size.
The latter
not matter
much
for whole countries, because cash flows between the players
converge
Table 3 to zero if the observed player pool are large (without rake, poker is a zero sum game).
Table 3 shows the ten biggest poker markets in the world in dollar terms of their gross market size and
Number of active online poker players, market share and market size per country
their share of the total market. Despite the prohibition of online gambling and the introduction of the
Unlawful Internet Gambling Enforcement Act (UIGEA) in
2006 the US was still by far the biggest
Gross Market size 2010
Rank
Country
in mil.
US$ per
year
Share
market with 973.3 million US$ in 2010. After theSize
Black
Friday,
when
the US sites of Pokerstars,
Full Tilt
26.95%
10.85%
2
Germany
391.94
The remaining
operatorsRussia
like Cake Poker or Bodog Poker235.12
which still accept US-players
had a market
6.51%
3
6.08%
4
Canada
219.63
share of less that 10% before the Black Friday. It will be interesting to see if these operators now rise and
5.19%
5
France
187.35
which level the US market will settle on.
4.42%
6
Great Britain
159.72
4.23%
7
Netherlands
152.80
The8 second largest market
of 392 million US$. Other
big markets are
3.24%
Spain is Germany with a market size
117.07
2.75%
99.25 market may now look different after
France,9Russia, Canada Sweden
and Great Britain. Although the French
2.24%
10
Finland
80.93
100%
Total
3,611.59
1
USA
973.30
Poker Absolute
Poker and
Ultimate Bet were shut down, the
US market size has dropped dramatically.
10
For a more detailed description of this procedure, see main report of the research project: Fiedler/Wilcke, 2011.
Prevalence
of from
Online
Poker
Our data were
09/09
to 03/10, but we assumed that these data are representative for 2010. Probably the
11
market grew a little in this time, but we neglected this and conservatively decided that it did not so we can give a
number
for the
Besides
thewhole
marketyear
size2010.
the prevalence of online poker is a key figure for describing the market for
12
According to PokerScout, the revenue of a poker site approximately consists of 70% cash games and 30% poker
online poker in a country. Prevalence can be measured in absolute terms (how many people play poker in
tournaments.
13
assumes
that
the share
of(how
revenues
the operators
is 70%
in every
a This
country)
and in
relative
terms
manyofpeople
play poker
in relation
to,country.
for example, the population).
14
The exact amount differs from operator to operator and also depends on the playing volume of the player. The
higher the playing volume, the higher the rakeback.
We derived the number of poker players in a country as follows: First, we used the country data 15 for
Gaming
Research identities
& Review
Journal ♦16Volume
Then, we16
the observed player identities UNLV
and added
the non identifiable
proportionally.
Issue 1
extrapolated the number by the market share of the observed sites during the data collection period. In a
last step we calculated the number of active players by assuming that per 100 player identities 85 real
persons exist (see reasoning above). The market shares are derived from the players in a country in
relation to the total number of players.
11
7 Prevalence of Online Poker
Besides the market size the prevalence of online poker is a key figure for describing
the market for online poker in a country. Prevalence can be measured in
absolute terms (how many people play poker in a country) and in relative
. . . there is a huge growth
terms (how many people play poker in relation to, for example, the
potential of the US market
population).
We derived the number of poker players in a country as follows: First,
should online poker be
we used the country data for the observed player identities and added
legalized.
the non identifiable identities proportionally. Then, we extrapolated
the number by the market share of the observed sites during the data
collection period. In a last step we calculated the number of active players by assuming
that per 100 player identities 85 real persons exist (see reasoning above). The market
shares are derived from the players in a country in relation to the total number of
players.
Table
4 lists
the number license,
of active pokerand
players
for the
tenthe
countriesmarkets
with theinmost
online
poker
online
poker
insulated
from
2010.
In total,
Table
4 was
listslegalized,
the number oftaxed,
active poker
players
for other
the ten countries
with
the the
most
players
and
their
share
of
all
players
in
the
world.
Concerning
the
market
size
the
USA
is
by
far
the
players paid
3.6players
billion US$
to the
operators
2010 or in
599the
US$
per player.
online
poker
andrake
their
share
of allinplayers
world.
Concerning the market
biggest
people
played
onlinethan
poker
formillion
real money
duringplayed
our dataonline
size
themarket.
USA More
is bythan
far 1.4
themillion
biggest
market.
More
1.4
people
Table 3forperiod
poker
real money
during
ourNearly
data collection
fromis09/09
to 03/10.
Nearly
collection
from 09/09
to 03/10.
every fourthperiod
poker player
an American
(23.71%).
On the
every
fourth
poker
player
is
an
American
(23.71%).
On
the
second
place
is
the
German
second place is the German market with 581 thousand players, or 9.64% of all players worldwide. On the
market
with
581
thousand
or7.39%),
9.64%
of all(401,701
players
worldwide.
On
the next ranks
Number
active
online
pokerplayers,
players,
share
market
size per
country
next
ranksoffollow
France
(445,860
players, market
Russiaand
players,
6.66%)
and Canada
follow France (445,860 players, 7.39%), Russia (401,701 players, 6.66%) and Canada
(401,701 players, 5.74%). Together, the ten countries with the highest number of poker players account
(401,701
players, 5.74%). Together, the ten countries with the highest number of poker
Gross Market size 2010
for 73.92%
of all players
worldwide.
players
account
for 73.92%
Rank
Country of all players worldwide.
Size in mil. US$ per year
Table 14
USA
973.30
2
Germany
391.94
3
Russia
10 countries with the highest number of active poker players 235.12
4
Canada
219.63
5
France
187.35
Rank
Country
Active players
6
Great
Britain
159.72
71
Netherlands
152.80
USA
1,429,943
82
Spain
117.07
Germany
581,350
93
Sweden
99.25
France
445,860
104
Finland
80.93
Russia
401,701
Total
3,611.59
5
Canada
345,971
Share
26.95%
10.85%
6.51%
6.08%
5.19%
4.42%
Share
4.23%
23.71%
3.24%
9.64%
2.75%
7.39%
2.24%
6.66%
100%
5.74%
4.47%
6
Great Britain
269,247
4.20%
7
Spain
253,043
Prevalence of Online Poker
3.98%
8
Netherlands
239,700
2.55%
9
Brazil
153,889
Besides the market size the prevalence of online poker is a key figure for describing the market for
2.15%
10
Australia
129,714
online poker in a country. Prevalence
can be measured in absolute
play poker in
Other
1,571,389terms (how many people26.06%
TOTAL
5,490,908
a country) and in relative terms
(how many people play poker
in relation to, for example, the100%
population).
15
We
the number
of poker
inpoker
a country
as follows:
First,
we used the country
dataof
for
Thederived
absolute
prevalence
of players
online
disregards
the
information
the size
a
The
absolute
prevalence
of online
poker disregards
the information
of the size of aof
country.
To know
16
Then,
wecountry,
the observed
player
identities
and added
the non identifiable
identities
proportionally.
country.
To
know
how
large
the
proportion
of
online
poker
players
is
in
a
given
how large the proportion of online poker players is in a given country, the number of players has to be
the
numberthe
ofnumber
players
to be related
population
of the
But period.
as not In a
extrapolated
byhas
the market
share of to
thethe
observed
sites during
the country.
data collection
related to the population of the country. But as not everybody has access to the internet, it is more
everybody
has
access
to
the
internet,
it
is
more
meaningful
to
relate
the
number
the
last step we calculated the number of active players by assuming that per 100 player identities 85to
real
meaningful
relate theusers
number
numberTable
of internet
users inthe
a country.
Table of
5 depicts
number
oftointernet
in toa the
country.
5 depicts
proportion
onlinethepoker
persons exist (see reasoning above). The market shares are derived from the players in a country in
proportioninofaonline
poker
in a country
per internet
0.307%
326) ofworldwide
all people
players
country
perplayers
internet
users. 0.307%
(1users.
out of
326) (1
ofout
allofpeople
relation to the total number of players.
with
internet
access
have
played
online
poker
for
real
money
during
the
6
months
worldwide with internet access have played online poker for real money during the 6 months of the of
data
collection.
Accounting
for
the
information
of the internet users in a country changes the ranking of the
15
For the players on Pokerstars and Cake Poker we obtained the city of origin. We used the city data base „World
countries.
thetocountries
withplayers
more to
than
100,000 internet users the proportion of online poker
Cities“
fromAmong
MaxMind
assign these
countries.
16
For 11.38% of all observed player identities we could not determine their country of origin. Adding them
players is highest for Hungary with 1.983%. One out of 50 Hungarians with internet access plays online
proportionally to the observed player identities of each country implies that they are equally distributed among all
countries.
may not For
be true
people
countries
stricter
laws
against
onlineingambling
a higher
poker for This
real money.
the as
USA,
it isfrom
striking
that itwith
is by
far the
biggest
market
absolutehave
terms,
but
incentive to hide their identity.
12
UNLV Gaming Research & Review Journal ♦ Volume 16 Issue 1
9 8 the data collection. Accounting for the information of the internet users in a country
changes the ranking of the countries. Among the countries with more than 100,000
th
internet
users
the proportion
of online
poker
is highest
forof Hungary
with
1.983%.
rank
only 36
considering
the number
of internet
usersplayers
in a country.
0.596%
all American
internet
th Hungarians with internet access plays online poker for real money. For
One
out
of3650
rank
only
considering the number of internet users in a country. 0.596% of all American internet
users (1 out of 168) are online poker players. On the one hand that means that the restrictions of the
theusers
USA,
it is
that poker
it is by
far the
biggest
market
in absolute
terms, but
rank
(1 out
of striking
168) are online
players.
On the
one hand
that means
that the restrictions
of the
th
UIGEA
were
effective
to the
some
extentofeven
before
Black
Friday.
On of
theallother
handof
that
means
that
rank only
36considering
considering
number
internet
usersthe
in users
a country.
0.596%
American
internet
only
36th
the
number
of
internet
in
a
country.
0.596%
all
American
th
UIGEA
were
effective tothesome
extent
before
theinBlack
Friday.
On theofother
hand thatinternet
means that
rank
only 36
considering
number
of even
internet
users
a country.
0.596%
all American
there
huge
growth
ofare
the
US
market
should
online
poker
be one
legalized.
usersis(1aout
of 168)
online
poker
players.
On the
one
hand
that means
that
the
restrictions
of means
the
internet
users
(1are
outpotential
of 168)
online
poker
players.
On the
hand that
that the
there
huge
growth
potential
ofplayers.
the US market
online
poker be
legalized.
users
(1 is
outa of
168)
are online
poker
On the should
one hand
that means
that
the restrictions of the
restrictions
of the UIGEA
wereeven
effective
toBlack
some
extent
before
themeans
Blackthat
Friday. On
UIGEA were effective
to some extent
before the
Friday.
On even
the other
hand that
UIGEA
effective
tomeans
some extent
even
before
the
Black
Friday. potential
On the otherofhand
that should
Table
5 awere
the
other
hand
thatpotential
there
is ashould
huge
growth
thethat
USmeans
market
there
is
growth
ofthat
the US
market
online
poker be legalized.
Table
5huge
there is apoker
huge growth
potential of the US market should online poker be legalized.
online
be legalized.
Table
5
10
countries
with
thethe
highest
pokerplayers
playersper
perinternet
internet
users
(only
countries
10 countries
with
highestproportion
proportion of
of online
online poker
users
(only
countries
withwith
Table 5
more
than
100,000
than
100,000
internetusers)
users)inin2010
2010
10 more
countries
with
theinternet
highest
proportion
of online poker players per internet users (only countries with
10 countries with the highest proportion of online poker players per internet users (only countries with
more
than 100,000
internet users)
in 2010of internet users in a country. 0.596% of all American internet
rank
only
36thCountry
considering
the number
Rank
Country
Active
players
Internet
Players/internet
useruser
Rank
Internetuser
user
Players/internet
more
than 100,000
internet users) inActive
2010 players
Hungary
122,482On the one hand
6,176,400
1.983%
users
168)
are online poker players.
that means that the restrictions
of the
1 1(1 out ofHungary
122,482
6,176,400
1.983%
Rank
Active players
Internet
user
Players/internet
user
Estonia
19,212
969,700
1.981%
2 2 were Country
Estonia
19,212
969,700
1.981%
UIGEA
effective
to some extent
even
before the Black
Friday.
other hand that
means that
Rank
Country
Active
players
Internet
user On the Players/internet
user
1 3
Hungary
122,482
6,176,400
1.983%
Portugal
100,075
5,168,800
1.936%
31
Portugal
100,075
5,168,800
1.936%
122,482
6,176,400
1.983%
there
growth
potential
of
the
US
market
should
online
poker
be
legalized.
4 a hugeHungary
2 is
Denmark
90,532
4,750,500
1.906%
Estonia
19,212
969,700
1.981%
42
Denmark
90,532
4,750,500
1.906%
Estonia
19,212
969,700
1.981%
3 5
Iceland
4,996
301,600
1.657%
Portugal
100,075
5,168,800
1.936%
53
Iceland
4,996
301,600
1.657%
Portugal
100,075
5,168,800
1.936%
4 65
Netherlands
239,700
14,872,200
1.612%
Denmark
90,532
4,750,500
1.906%
Table
654 7
Netherlands
239,700
14,872,200
1.612%
Denmark
90,532
4,750,500
1.906%
Finland
71,543
4,480,900
1.597%
Iceland
4,996
301,600
1.657%
765 8
Finland
71,543
4,480,900
1.597%
Iceland
4,996
301,600
1.657%
Netherlands
239,700
14,872,200
1.612%
Cyprus
6,445
433,800
1.486%
10876countries
with
the highest proportion
of online poker 14,872,200
players
per internet users1.597%
(only
countries with
Netherlands
239,700
1.612%
Cyprus
6,445
433,800
1.486%
9
Norway
64,535
4,431,100
1.456%
Finland
71,543
4,480,900
Finland
71,543
4,480,900
1.597%
987 10
Norway
64,535
4,431,100
1.456%
Slovenia
18.899
1,298,500
1.455%
Cyprus
6,445
433,800
1.486%
more
than 100,000
internet users) in6,445
2010
Cyprus
433,800
1.486%
10
98
Slovenia
18.899
1,298,500
1.455%
…
…
…
…
Norway
64,535
4,431,100
1.456%
9 36
Norway
64,535
4,431,100
1.456%
10
USA
1,429,943
239,893,600
0.596%
Slovenia
18.899
1,298,500
1.455%
…
…
…
…
10
Slovenia
18.899
1,298,500
1.455%
……
……
… … user
……
36
Rank
Country
Active
players
Internet
Players/internet
USA
1,429,943
239,893,600
0.596% user
…
……
… …
…
361
TOTAL
6,029,930
1,965,162,316
0.307%
USA
1,429,943
239,893,600
0.596%
…
…
Hungary
122,482
6,176,400
1.983%
36 internet user:
USA
239,893,600
0.596%
Source
Internet World Stats,1,429,943
for June
2010.
…
…
…
…
TOTAL
6,029,930
1,965,162,316
0.307%
2
Estonia
19,212
969,700
1.981%
…
…
…
…
TOTAL
6,029,930
1,965,162,316
0.307%
Source3internet user:
Internet World Stats, for
June
2010.
Portugal
100,075
5,168,800
1.936%
Drivers
ofTOTAL
the
prevalence
of6,029,930
Online
1,965,162,316
0.307%
Source
internet user:
Internet
World Stats, for
June
2010. Poker
Source4 internet user:Denmark
Internet World Stats, for June 90,532
2010.
4,750,500
1.906%
Drivers
of theIceland
prevalence of Online
Poker
5 To understand
4,996
Drivers
of
Online
Poker
Drivers
of the
theprevalence
prevalence
Online
Poker
the drivers of
ofof
the
market
for
online poker, it is301,600
most useful to analyze the 1.657%
relative size
Drivers
of
the
prevalence
of
Online
Poker
6
Netherlands
239,700
14,872,200
1.612%
To
understand
the
drivers
of
the
market
for
online
poker,
it
is
most
useful
toonanalyze
of
the markets because
the absolute
numbersfor
mostly
depend
on the
size ofuseful
a country.
We focused
the size
To
thedrivers
drivers
themarket
market
online
poker,
is most
to analyze
the relative
7 understand
Finland
71,543
4,480,900
1.597%
To
understand
the
ofofthe
for online
poker,
it is itmost
useful tomostly
analyze
the relative
size
therelative
relative
size
of
the
markets
because
the
absolute
numbers
depend
on
the
size
prevalence
of
online
poker
and
identified
that
the
two
main
drivers
for
this
number
are
GDP
per
To
understand
the drivers
of the market
for online
poker,
it ison
most
useful
to aanalyze
theWe
relative
size on the
8markets
Cyprus
6,445
1.486%
of
because
theabsolute
absolute
numbers
mostly
depend
the
size
of
country.
focused
ofthe
the
markets because
the
mostly
depend
on the
size
of
a country.
We
on the that
of
a country.
We
focused
on numbers
the
relative
prevalence
of433,800
online
poker
andfocused
identified
9 markets
capita
and culture.
Athe
simple
regression
model
withdepend
the proportion
of of
poker
playersWe
perfocused
internet
users
Norway
64,535
1.456%
of the
because
absolute
numbers
mostly
on the4,431,100
size
a country.
on
the per
relative
ofonline
online
poker
and
identified
that
two
main
drivers
fornumber
thisA
number
areper
GDP per
the
twoprevalence
main Slovenia
drivers
forpoker
thisand
number
arethat
GDP
per
capita
and
simple
regression
relative
of
identified
the the
two
main
drivers
forculture.
this
are
GDP
10 prevalence
18.899
1,298,500
country
and its GDP
per capita
yielded
a significant
influence:
If
the GDP
per
capita
risesare
by1.455%
1,000per
US$,
relative
prevalence
of
online
poker
and
identified
that
the
two
main
drivers
for
this
number
GDP
model
with
the
proportion
of
poker
players
per
internet
users
per
country
and
its
GDPper
capita
simpleregression
regression
model
with
proportion
of poker
players
per internet
users
capita and
and culture.
culture. AA
model
the the
proportion
of poker
players
per internet
users
…simple
…with
…
… per
17
the
proportion
of
online
poker
players
increases
by
0.009
percentage
points
(t=5.786,
p<.001).
capita
and culture.
A simple
regression influence:
model with the
proportion
ofper
poker
players
per internet
usersUS$,
per
per
capita
yielded
a
significant
If
the
GDP
capita
rises
by
1,000
36 and
USA
0.596%
country
and its GDP
per
a1,429,943
significant
influence:
If 239,893,600
theIfGDP
per capita
rises by
1,000
country
GDP
percapita
capitayielded
yielded
a significant
influence:
the GDP
per capita
rises
by US$,
1,000 US$,
country
and its GDP
capitapoker
yielded players
a significant
influence:by
If the
GDP
per capita risespoints
by 1,000
the
proportion
of…per
online
increases
0.009
17 (t=5.786,
… by
… percentage
…US$,
17
the
proportion
of
online
poker
players
increases
0.009
percentage
points
(t=5.786,
p<.001).
Table 6
the proportion
of online poker players increases by 0.009 percentage points (t=5.786, p<.001).
p<.001).
6,029,930
1,965,162,316
0.307%
the proportion ofTOTAL
online poker players increases
by 0.009 percentage
points (t=5.786, p<.001).17
Source internet user: Internet World Stats, for June 2010.
Table 6
Table
6 6 of
a simple
linear
regression
Table
Results
regarding the influence of GDP per capita on the relative
17
Drivers
of theforprevalence
of Online
Poker
See of
appendix
the operationalization
of the simple
regression.
Results
a simple
linear
regression regarding
thelinear
influence
of GDP per capita on the relative
of
online
poker
prevalence
10 17 See
To
appendix
understand
for
the
the
operationalization
drivers
of
theofmarket
the simple
for linear
onlineregression.
poker, it is most useful to analyze the relative size
prevalence
of online poker
17
17 See
for the
theoperationalization
operationalization
simple
linear
regression.
Seeappendix
appendix for
of of
thethe
simple
linear
regression.
of the markets because the absolute numbers mostly depend on the size of a country. We focused 10 on the
Sample (n=161)
10 10 relative
prevalence
of
online
poker
and
identified
that
the
two
main
drivers
for
this
number
are
GDP
per
Sample (n=161)
Regression
capita and culture. A simple regression
model with the proportion
internet users per
Variable
t-valueof poker players perSignificance
coefficient
Regression
Variable
t-value
Significance
coefficient
country and its GDP per capita yielded a significant influence: If the GDP per capita rises by 1,000 US$,
Constant
.192
4.123
.000
the proportion of online poker players increases by 0.009 percentage points (t=5.786, p<.001).17
GDP
per capita in
Constant
.192
4.123
.000
1000
.009
5.786
.000
GDP US$
per capita in
Table
6
1000
US$
.009
5.786
.000
Goodness of fit
R² = .177; adjusted R² = .171; F-value= 33.477 (p<.000)
Goodness of fit
R² = .177; adjusted R² = .171; F-value= 33.477 (p<.000)
17
18
See appendix
foralso
the operationalization
of thethat
simple
regression.
We were
able to find evidence
thelinear
culture
has a significant influence on the proportion
18
We
were
also
able
to
find
evidence
that
the
culture
has
aJournal
significant
influence
the 1proportion
13
Research
& Review
♦ Volume
16 on
Issue
of poker players per internetUNLV
users Gaming
in a country.
An analysis
of
variance
(ANOVA)
with
a F-Value
of 10 of poker(p<.001)
players confirmed
per internetthat
users
in aare
country.
An analysis
of variance
(ANOVA) with
a F-Value
14.114
there
significant
differences
in the mean-values
of each
culturalofgroup
14.114
(p<.001)
confirmed
that
there
are
significant
differences
in
the
mean-values
of
each
cultural
group
(see table 7). We controlled the effect for GDP per capita by using it as a covariate (F =12.753,
p<.001).
(see
7). Weand
controlled
the effect
for GDPof
perthecapita
by using
it as the
a covariate
=12.753,
p<.001).
GDPtable
per capita
culture explain
R²=50.4%
variation
between
relative (F
poker
prevalence
in
Constant
GDP per capita in
1000 US$
Goodness of fit
.192
4.123
.000
.009
5.786
.000
R² = .177; adjusted R² = .171; F-value= 33.477 (p<.000)
18
were
alsoable
able to
to find
thatthat
the culture
has a has
significant
influenceinfluence
on the proportion
WeWe
were
also
findevidence
evidence
the culture
a significant
on
of poker
players per
users in aper
country.
An analysis
of variance
(ANOVA)
with a F-Value
of
the
proportion
of internet
poker players
internet
users in
a country.
An analysis
of variance
(ANOVA)
with
a
F-Value
of
14.114
(p<.001)
confirmed
that
there
are
significant
14.114 (p<.001) confirmed that there are significant differences in the mean-values of each cultural group
differences
in the mean-values of each cultural group (see table 7). We controlled the
(see table 7). We controlled the effect for GDP per capita by using it as a covariate (F =12.753, p<.001).
effect for GDP per capita by using it as a covariate (F =12.753, p<.001). GDP per capita
GDP per capita and culture explain R²=50.4% of the variation between the relative poker prevalence in
and culture explain R²=50.4% of the variation between the relative poker prevalence in
the countries.
the
countries.
Table 7
Results of the ANOVA regarding the influence of culture on the relative prevalence of online poker
Sample (n=161)
Sum of
Results of a simple linear regression
GDP per capita on
Squares regarding
dfthe influence
MeanofSquare
F the relative
Sig.
prevalence of online poker
Between Groups
17.974
8
2.247
Constant
2.136
1
2.136
GDP per capita
1.483
1
Sample
(n=161)1.483
Cultural group
11.512 Regression7
1.645
Within
17.673 coefficient
152
.116
VariableGroups
t-value
TOTAL
35.647
161
Constant
GDP
Fit of per
the capita
modelin
1000 US$
.192
4.123
R² = .504; adjusted R² = .478
.009
5.786
19.324
18.375
12.753
14.114
.000
.000
.000
.000
Significance
.000
.000
The results
implyonline
that onlineis poker
is only a factor in western and orthodox
only a factor in western and orthodox countries. With the
Goodness
The
results
of
fit
imply
that
poker
R² = .177;
adjusted R²from
= .171;
F-value=
33.477
(p<.000)
countries.
With
the
exception
of
the
countries
Latin
America,
markets in all
18
We operationalized
culture
to the the
classification
1996:the
1=Western,
exception
of the countries
fromaccording
Latin America,
markets inofallHuntington
other countries
are
negligible2=Orthodox,
(see table
other
countries
are
(see
table7=Hindu,
8). That8=Buddhist,
can either
mean see
that
these markets have
3=Islamic,
4=African,
5= negligible
Latin American,
6=Sinic,
0=Others;
appendix.
8). That
can either mean that
these markets
an enormous
potential
or that
of these
18 growth
an
enormous
or thathave
people
of these
groups
arepeople
nottheinterested
We weregrowth
also ablepotential
to find evidence
that
the culture
has cultural
a significant
influence
on
proportion
11 cultural
groups
are
not
interested
in
online
poker
at
all.
The
directions
of
these
results
could
be
confirmed
in
online
poker
at
all.
The
directions
of
these
results
could
be
confirmed
in
a
Turkey-test
poker players per internet users in a country. An analysis of variance (ANOVA) with a F-Value
of
of
which
compares
thecompares
mean values
(for
detailed
information
see see
table
1212ininthe
in a Turkey-test
which
the mean
values
(for detailed
information
table
the appendix).
appendix).
14.114 (p<.001) confirmed that there are significant differences in the mean-values of each cultural group
(see table 7). We controlled the effect for GDP per capita by using it as a covariate (F =12.753, p<.001).
Table 8
GDP per capita and culture explain R²=50.4% of the variation between the relative poker prevalence in
the
countries.statistics of the ANOVA regarding the influence of culture on the relative prevalence of
Descriptive
online7poker
Table
(n=161)
Results of the ANOVA regardingSample
the influence
of culture on the relative prevalence of online poker
95%-confidence
interval
Sample
(n=161)
Lower
Upper
Standard
Standard
of
Culture
N
Mean Sumdeviation
bound
bound
error
Squares
df
Mean Square
F
Sig.
Western
47
.638
.452
.040
.558
.718
OrthodoxGroups
13
.450 17.974.326
.348
.651 19.324
Between
8 .077
2.247
.000
Islamic
41
.076
.108
.043
-.009
.161
Constant
2.136
1
2.136
18.375
.000
African
23
.079
.059
.058
-.035
.193
GDP
per
capita
1.483
1
1.483
12.753
.000
Latin
Cultural
7 .062
1.645
.000
Americangroup
20
.185 11.512.100
.063
.308 14.114
Sinic Groups
6
.019 17.673.018
-.204
.242
Within
152.113
.116
Hindu
5
.112 35.647.196
-.132
.356
TOTAL
161.124
Buddhist
6
.128
.157
.113
-.095
.351
Fit
of the model
TOTAL
161
.289
.369
R² = .504; adjusted R² = .478
An
analysis
of
the
influence
of
the law on the prevalence on online poker yielded no significant
18
We operationalized culture according to the classification of Huntington 1996: 1=Western, 2=Orthodox,
3=Islamic,
4=African,
5= the
Latin
American, 6=Sinic,
7=Hindu,
0=Others;
see appendix.
results. But
as law and
enforcement
of law are
hard to8=Buddhist,
quantify this
result may
not be taken for full
value. In our
eyes Gaming
it mainly Research
says that law
which prohibits
poker16
was
not enforced
at the time of the
14
11 UNLV
& Review
Journalonline
♦ Volume
Issue
1
data collection.
Limitations
An analysis of the influence of the law on the prevalence on online poker yielded no
significant results. But as law and the enforcement of law are hard to quantify this result
may not be taken for full value. In our eyes it mainly says that law which prohibits online
poker was not enforced at the time of the data collection.
Limitations
Although the study yields many findings there are some limitations. First, we
only collected data about cash games, poker tournaments have not been taken into
consideration. As mentioned in footnote 12 the revenue of a poker site approximately
consists of 70% cash games and 30% poker tournaments. The market size calculations of
this study assume that this relation is identical for every operator, which might not be true
as the tournaments require a larger player pool and the smaller poker networks probably
earn a lesser percentage with them. The OPD-UHH also excludes pure tournament
players and, hence, the numbers of poker players in the different countries we derived in
The
calculations ofthe
thistrue
studyvalue.
assumeHowever,
that this relation
is identical for
every
operator,
which
thismarket
studysize
underestimate
it is reasonable
that
there
are just
a very
might
not be true
as the tournaments
require a larger
player pool
andnot
the smaller
poker
networks
few players
playing
only tournaments
and whom
we do
have any
data
on. probably
the datawith
on the
origin
is self reported
by thepure
players.
It isplayers
therefore
earn Second,
a lesser percentage
them.
The OPD-UHH
also excludes
tournament
and, possible
hence, the
that
not
all
players
stated
their
correct
country.
But
as
withdrawing
funds
is
only
numbers of poker players in the different countries we derived in this study underestimate the true possible
value.
if the correct personal data are submitted, we reason that the cases of wrong information
However, it is reasonable that there are just a very few players playing only tournaments and whom we do
are rare.
not have any data on.
Third, as PokerStars and Cake Poker do not give information about the country of a
player but only of their city. To assign the players to countries we used the city database
Second, the data on the origin is self reported by the players. It is therefore possible that not all players
“World Cities” by Maxmind. In doing so we faced two general challenges: 1) Many city
stated their correct country. But as withdrawing funds is only possible if the correct personal data are
names in the world exist multiple times. 2) The city data base only lists cities having a
submitted,
weof
reason
that the
cases of
are rare.
population
at least
50,000.
Towrong
solveinformation
these issues
we developed an algorithm to assign
player identities to countries by taking the size of the cities into consideration. In the end,
PokerStars and Cake Poker do not give information about the country of a player but only of
weThird,
wereasable
to successfully assign 92% of the poker players stating a city of origin to
their
To assign
the players
countries we
city database “World Cities” by Maxmind.19 In
theircity.
country
of origin.
8%toremained
asused
not the
assignable.
Finally,
we faced
the problem
one player
caninhave
moreexist
than
one poker
doing
so we faced
two general
challenges:that
1) Many
city names
the world
multiple
times. identity
2) The
by
registering
at
multiple
sites.
It
may
also
be
possible
that
more
than
one
person
use the
city data base only lists cities having a population of at least 50,000. To solve these issues we developed
same
player
identity
(for
example
family
members).
Considering
different
qualitative
an algorithm to assign player identities to countries by taking the size of the cities into consideration. In
arguments (see footnote 8), we estimate that for 100 poker identities 85 real persons
the end, we were able to successfully assign 92% of the poker players stating a city of origin to their
exist. We admit this number is not an exact empirical value but an estimation and
country
of origin.
remained as interpretation.
not assignable.
therefore
open 8%
to subjective
Finally, we faced the problem that one player can have more than one poker identity by registering at
Summary and Perspectives
multiple
It may
also market
be possible
more rapidly
than one person
same player
identity
(forthere
example
Thesites.
online
poker
hasthat
grown
in the use
pasttheyears.
But until
now
family
members).
qualitative
arguments
8),knew
we estimate
that for 100
have only
beenConsidering
estimates different
about the
total market
size(see
andfootnote
nobody
how where
the
players
and their
money
which
countries
are
markets,
poker
identities
85 real
personscome
exist. from,
We admit
this number
is not
an the
exactbiggest
empirical
value butand
an
where
the
proportion
of
people
gambling
online
on
poker
games
in
relation
to
the
internet
users
estimation and therefore open to subjective interpretation.
is highest. We were first to answer these questions by using data from the Online Poker
Database of the University of Hamburg (OPD-UHH) which includes information about
approximately 4.6 million online
poker players
and their countries of origin.
Summary
and Perspectives
This study is able to shed light on the online poker market. It is the first time that
The online poker market has grown rapidly in the past years. But until now there have only been
the market can be broken down to countries which provides important insights for local
estimates
about
the total
market sizewant
and nobody
knew how
where
the players
their gambling.
money comeWe
legislators
who,
for example,
to evaluate
their
regulation
ofand
online
from,
whichan
countries
are the
markets, of
andthe
where
the proportion
of the
people
gambling
online
first gave
overview
ofbiggest
the structure
online
poker and
market
shares
ofon
the
different
poker
sitestointhethe
yearsusers
2008-2010.
described
the
data
in the OPDpoker
games
in relation
internet
is highest.We
We then
were first
to answer
these
questions
by using
UHH
and
methodology
of gathering
it. Weofwere
able(OPD-UHH)
to identifywhich
4,591,298
data
from
thethe
Online
Poker Database
of the University
Hamburg
includespoker
identities
and
extrapolated
this
number
to
6,029,980
different
players
who
paid 3.6
information about approximately 4.6 million online poker players and their countries of origin.
billion US$ rake to the operators in 2010 (on average 599 US$ per player). Hence, online
poker has evolved to become a huge market. This is especially true for western and
19
For further information about this database see: http://www.maxmind.com/app/worldcities.
UNLV Gaming Research & Review Journal ♦ Volume 16 Issue 1
15
13 orthodox countries where most of the business takes place – even when controlling for
GDP – per capita.
We also assigned the observed poker identities to their country of origin and derived
the market size of the most important poker countries: Even though the USA had been
the biggest market with 1.4 million players paying 973 million US$ rake in 2010 (before
Black Friday), there are 35 countries with a higher prevalence of online poker in relation
to the number of internet users in that country. For example, Hungary has the highest
proportion of online poker players in relation to the number of internet users (1.98%).
The relatively low prevalence of players in the USA could be seen as evidence that
the regulation of online poker and the UIGEA deterred at least some
of the potential players even before Black Friday. On the other hand
But the most important cause
this comparison with the prevalence of online poker in other countries
for a regulated market being
means that there is a lot of potential in the American online poker
market. However, these numbers are mostly generated from totally free
a smaller market is a fenced
(black) markets. However, a regulated market will probably not yield
market, which excludes nonthe full potential of the market. One reason is taxes, which increase the
residents from the player pool
rake, make the product more expensive, and lead to reduced demand.
Another reason might be player protection which helps addicted
(e.g. Italy or France).
gamblers to stay away from the game as addicts tend to play longer
sessions, more often and more intensely than non-addicts (e.g. see
Productivity Commission 2010) and, hence, are an important source of
revenue (Williams & Wood 2004). But the most important cause for a regulated market
being a smaller market is a fenced market, which excludes non-residents from the player
pool (e.g. Italy or France). This sharply reduces the player pools and, hence, the positive
network effects of large player pools where it is possible to find enough people to play
with for every game type, every limit and at any daytime.
The results presented in this paper did not include information on the market after
the Black Friday when the most important poker operators in the USA were shut down.
Further research has to determine the effects of this event and clarify if the players from
the USA stopped playing, switched to other operators still accepting US players, or even
avoid the country identification and log in to Pokerstars or Full Tilt Poker with a proxyIP and deposit with an international credit card.
This study only scratches on the surface of the possibilities the detailed data in
the ODP-UHH allow. It is possible not only to break down the market on a national
level but also to regions within countries. Future studies will also be able to analyze
the playing habits and to determine for example the session lengths and the playing
frequency of a player and which limits and how many tables the player plays
simultaneously. This information can then be used for further analyses regarding inter
and intra country differences. By conducting time series analyses it is also be possible
to identify how many players tend to play more often and more intensively over time
and how many decrease the intensity of play and what kind of factors may give a hint at
which group a player belongs to. This may not only be interesting for the industry but
also from the viewpoint of excessive and compulsive gambling.
16
UNLV Gaming Research & Review Journal ♦ Volume 16 Issue 1
play more often and more intensively over time and how many decrease the intensity of play and what
kind of factors may give a hint at which group a player belongs to. This may not only be interesting for
the industry but also from the viewpoint of excessive and compulsive gambling.
Measurement Appendix
Measurement Appendix
Table 9
Operationalization of s variables of the simple linear regression regarding relative market size and GDP
per capita
Variable
Description
independent variable
in thousand US-Dollar
GDP per Capita
Proportion of poker players
dependent variable
proportion of poker players in a country
per internet users in percentage
Table 10
Operationalization of variables of ANOVA
Variable
Description
dependent variable
proportion of poker players in a country
per internet users in percentage
Proportion of poker players
factor
1=Western, 2=Orthodox, 3=Islamic,
4=African, 5= Latin American, 6=Sinic,
7=Hindu, 8=Buddhist, 0=Others
Cultural group
Covariate
In thousand US-Dollar
GDP per capita
15 Table 11
Players per Stake for No Limit Holdem
Limit (Small Blind/Big Blind in $)
Stakes
% of players
0,01/0,02-0,05/0,10
0,10/0,20- 0,5/1
0,75/1,50-5/10
Micro
Low
Mid
48.43%
41.24%
9.97%
8/16-500/1000
High
0.36%
Table 12
Influence of Culture – Results of Turkey-Test
Culture (I)
Western
Orthodox
Sample (n=161)
Standard
p-value
error
Culture (J)
Average
difference (I-J)
Orthodox
Islamic
African
Latin American
Sinic
Hindu
Buddhist
Western
Islamic
African
Latin American
Sinic
Hindu
Buddhist
Western
Orthodox
.176
.111
.760
.717*
.076
.000
.713*
.090
.000
.577*
.094
.000
.789*
.153
.000
.670*
.166
.002
.649*
.153
.001
-.176
.111
.759
.541*
.113
.000
UNLV
Gaming
Research
&
.538*
.123
.001
.401*
.126
.037
.614*
.175
.013
.495
.186
.144
.474
.175
.127
-.717*
.076
.000
-.541*
.113
.000
95%-confidence interval
Lower bound Upper bound
-.165
.516
.484
.949
.436
.990
.287
.867
.318
1.261
.159
1.182
.178
1.121
-.516
.165
.195
.887
Review
Journal
♦ Volume
.160
.915
.014
.789
.077
1.150
-.077
1.067
-.063
1.011
-.949
-.484
-.887
-.195
16 Issue 1
17
0,01/0,02-0,05/0,10
0,10/0,20- 0,5/1
0,75/1,50-5/10
Micro
Low
Mid
48.43%
41.24%
9.97%
8/16-500/1000
High
0.36%
Table 12
Influence of Culture – Results of Turkey-Test
Culture (I)
Western
Orthodox
Islamic
African
LatinAmerican
Sinic
Hindu
Buddhist
Culture (J)
Average
difference (I-J)
Sample (n=161)
Standard
p-value
error
95%-confidence interval
Lower bound Upper bound
Orthodox
Islamic
African
Latin American
Sinic
Hindu
Buddhist
Western
Islamic
African
Latin American
Sinic
Hindu
Buddhist
Western
Orthodox
African
Latin American
Sinic
Hindu
Buddhist
Western
Orthodox
Islamic
Latin American
Sinic
Hindu
Buddhist
Western
Orthodox
Islamic
African
Sinic
Hindu
.176
.717*
.713*
.577*
.789*
.670*
.649*
-.176
.541*
.538*
.401*
.614*
.495
.474
-.717*
-.541*
-.003
-.140
.073
-.046
-.067
-.713*
-.538*
.003
-.136
.076
-.043
-.064
-.577*
-.401*
.140
.136
.212
.094
.111
.076
.090
.094
.153
.166
.153
.111
.113
.123
.126
.175
.186
.175
.076
.113
.092
.097
.155
.168
.155
.090
.123
.092
.108
.162
.175
.162
.094
.126
.097
.108
.165
.177
.760
.000
.000
.000
.000
.002
.001
.759
.000
.001
.037
.013
.144
.127
.000
.000
1.000
.833
1.000
1.000
1.000
.000
.001
1.000
.912
1.000
1.000
1.000
.000
.037
.833
.912
.902
.999
-.165
.484
.436
.287
.318
.159
.178
-.516
.195
.160
.014
.077
-.077
-.063
-.949
-.887
-.287
-.436
-.403
-.561
-.542
-.990
-.915
-.280
-.469
-.423
-.579
-.562
-.867
-.789
-.157
-.196
-.294
-.450
.516
.949
.990
.867
1.261
1.182
1.121
.165
.887
.915
.789
1.150
1.067
1.011
-.484
-.195
.280
.157
.548
.469
.408
-.436
-.160
.287
.196
.574
.494
.435
-.287
-.014
.436
.469
.718
.637
Buddhist
Western
Orthodox
Islamic
African
Latin American
Hindu
Buddhist
Western
Orthodox
Islamic
African
Latin American
Sinic
Buddhist
Western
Orthodox
Islamic
African
Latin American
Sinic
Hindu
.073
-.789*
-.614*
-.073
-.076
-.212
-.119
-.140
-.670*
-.495
.046
.043
-.094
.119
-.021
-.649*
-.474
.067
.064
-.073
.140
.021
.165
.153
.175
.155
.162
.165
.214
.204
.166
.186
.168
.175
.177
.214
.214
.153
.175
.155
.162
.165
.204
.214
1.000
.000
.013
1.000
1.000
.902
.999
.997
.002
.144
1.000
1.000
.999
.999
1.000
.001
.127
1.000
1.000
1.000
.997
1.000
-.434
-1.261
-1.150
-.548
-.574
-.718
-.777
-.768
-1.182
-1.067
-.469
-.494
-.637
-.540
-.679
-1.121
-1.011
-.408
-.435
-.579
-.488
-.638
.579
-.318
-.077
.403
.423
.294
.540
.488
-.159
.077
.561
.579
.450
.777
.638
-.178
.063
.542
.562
.434
.768
.679
Table 12: Results of Turkey-Test.
*: p < .05.
16 References
Cabot, A., & Hannum, R. (2008). Poker: Public Policy, Law, Mathematics and the Future of an American
Tradition. T.M. Cooley Law Review, 22, 443-513.
Chen, B., & Ankenman, J. (2006). The Mathematics of Poker. Pittsburgh : ConJelCo.
DeDonno, M.A., & Detterman, D.K. (2008). Poker Is a Skill. Gaming Law Review and Economics, 12,
31-36.
& Borm, P., & van der Genugten, B. (2003). On Strategy and Relative Skill in Poker.
18Dreef, M.,
UNLV Gaming
Research
& Review
Journal ♦ Volume 16 Issue 1
International
Game
Theory Review,
5, 83-103.
Fiedler, I.,& Wilcke, A.-C. (2012). Online Poker in the European Union. Gaming Law Review and
Economics, 16, 21-26.
Fiedler, I., & Wilcke, A.-C. (2011). Der Markt für Onlinepoker: Spielerherkunft und Spielerverhalten.
References
Cabot, A., & Hannum, R. (2008). Poker: Public Policy, Law, Mathematics and the Future
of an American Tradition. T.M. Cooley Law Review, 22, 443-513.
Chen, B., & Ankenman, J. (2006). The Mathematics of Poker. Pittsburgh : ConJelCo.
DeDonno, M.A., & Detterman, D.K. (2008). Poker Is a Skill. Gaming Law Review and
Economics, 12, 31-36.
Dreef, M., & Borm, P., & van der Genugten, B. (2003). On Strategy and Relative Skill in
Poker. International Game Theory Review, 5, 83-103.
Fiedler, I.,& Wilcke, A.-C. (2012). Online Poker in the European Union. Gaming Law
Review and Economics, 16, 21-26.
Fiedler, I., & Wilcke, A.-C. (2011). Der Markt für Onlinepoker: Spielerherkunft und
Spielerverhalten. Norderstedt: BOD Verlag,
Fiedler, I., & Rock, J.-P. (2009). Quantifying skill in games – Theory and empirical
evidence for poker. Gaming Law Review and Economics, 13, 50-57.
Haruyoshi, S. (2005). Internet Poker: Data Collection and Analysis, Brown University.
H2GC. (2009). eGambling data bulletin. 2008 Q3/Q4 report. H2 Gambling Capital.
Internet World Stats. (June 2010). http://www.internetworldstats.com/list2.htm.
Huntington, S. P. (1996). The Clash of Civilizations. New York: Simon & Schuster.
Partygaming, 2010: 2009 full year results presentation.
Productivity Commission 2010, Gambling - Productivity Commission Inquiry Report,
Australian Government.
Rock, J.-P., & Fiedler, I. (2008). Die Empirie des Online-Pokers – Bestimmung des
Geschicklichkeitsanteils anhand der kritischen Wiederholungshäufigkeit. Zeitschrift für
Wett- und Glücksspielrecht, 3, 412-422.
Turner, N. E. (2008). Viewpoint: Poker Is an Acquired Skill. Gaming Law Review and
Economics, 12. 229-30.
Williams, R. J., & Wood, R. T. (2004). The proportion of gaming revenue derived from
problem gamblers: Examining the issues in a Canadian context. Analyses of Social
Issues and Public Policy, 4,33–45.
UNLV Gaming Research & Review Journal ♦ Volume 16 Issue 1
19
20
UNLV Gaming Research & Review Journal ♦ Volume 16 Issue 1