When 4 million people worldwide checked into some of the world鈥檚 most popular poker websites and played an estimated billion hands, software was watching.
The results reveal patterns in play that could help inform how poker is regulated and uncover a wealth of information about one of the internet鈥檚 most popular pastimes. 鈥淚t鈥檚 a data gold mine,鈥� says economist at the University of Hamburg in Germany. 鈥淚t is interesting for regulators, academics and also for the treatment of problem gamblers.鈥�
Fiedler gathered his data on the 4 million players between September 2009 and March 2010. To do so, he turned to the poker-market spectator . Its software logged the , game outcomes, the date and time, and the commission paid to the operator by people playing on the two biggest sites worldwide at the time 鈥� Pokerstars and 鈥� as well as smaller operators, Everest Poker, IPN Poker and Cake Poker.
Advertisement
Altogether, that amounted to 4.6 million different online-poker 鈥渋dentities鈥� worldwide, which Fiedler reckons equates to about 3.9 million different players; some players play under different screen names on several different poker sites. As those websites accounted for about two-thirds of the market at the time, he used the data to extrapolate a figure for the total number of people playing online poker worldwide: 6 million. Click here to see the geographic spread of poker players and current legal status.
But this information, which first appeared in a book co-authored by Fiedler and released earlier this year in German called , also reveals patterns in play, painting a picture of a few dominant players who play a huge amount, and a majority who barely play at all. Fifty per cent of people played for less than 5 hours in a period of six months, while 6 per cent played for more than 100 hours. 鈥淰ery few people account for a lot of the playing volume,鈥� says Fiedler.
鈥淎 few dominant players play a huge amount, while more than half of players barely play at all鈥�
Although he did not capture the amount of money exchanged between players, Fiedler can still get some idea of . He recorded the average rake 鈥� the amount of money paid to the poker site by a player 鈥� per hour (Click here to see amount paid by players to online poker sites). Generally, the higher the stakes being played for, the higher the rake.
Fiedler says the next challenge is teasing apart which of the people who play intensively are pathological gamblers as opposed to professionals 鈥� perhaps by finding ways to use the data to measure their gambling patterns. Impulsive betting, for example, is one way to tell a problem gambler from a professional.
Kahlil Philander, who studies gambling policy at the University of Nevada in Las Vegas says that this could help further our understanding of gaming behaviour. 鈥淥nline poker is a relatively benign activity for 95 to 99 per cent of its users, but is very intense for a handful of professionals and potential pathological gamblers,鈥� he says. 鈥淔urther distinguishing between those two groups is the next challenge for player-analytics software, to help determine which players may be at-risk gamblers.鈥�
When this article was first posted, the first sentence read: 鈥淏ETWEEN September 2009 and March 2010, software watched as about 4 million people worldwide checked into some of the world鈥檚 most popular poker websites and played an estimated billion hands.鈥�
Poker: a game of skill for the few and of luck for the rest
ARE online poker players playing a game of skill or chance? It鈥檚 a contentious issue, as the so-called 鈥減redominance test鈥� 鈥� a measure of whether skill or chance is more dominant 鈥� is used in many US jurisdictions, and others across the world, to settle whether poker should be treated as gambling. In many places, this is what determines its legality.
In 2009, Ingo Fiedler and his colleague Jan-Philipp Rock at the University of Hamburg, Germany, reasoned that while each individual hand has elements of both skill and luck, over time winnings or losses due to chance should cancel out, whereas those due to skill, or a lack of it, should be consistent and accumulate. He tracked how 55,000 online players鈥� winnings and losses varied over time and calculated that a typical player, who loses overall, would have to play around 1560 hands, equivalent to about 22 hours online, before their winnings or losses were more a result of their skill 鈥� or lack of it 鈥� than luck. The pair called this point the 鈥渃ritical repetition frequency鈥� (CRF).
Now new figures reveal that most people play less than 5 hours over the course of 6 months 鈥� not enough for their skill level to become the dominant factor. On the other hand, Fiedler discovered that professional players 鈥� those who regularly win 鈥� have a much higher CRF. It turns out that far more hands need to be played for skill to show through than a lack of skill. That is because skilled players win less per hand than bad players lose. Or to put it another way, it is much easier to throw your money away than to win other people鈥檚.
Professionals play so many hands, often by playing on multiple tables at once, that they easily pass their CRF (click here to see how professional and typical players differ). 鈥淭hey are playing a game of skill,鈥� says Fiedler.
The predominance test is misguided because of this difference between the two types of players, he says. 鈥淭he legalisation of online poker should not depend on the degree of skill of the game, but the potential for addiction,鈥� he says, something he hopes the kind of data he has gathered will shed light on.
