Is poker a game of skill, or a game of luck?
Last week, researchers claimed to have developed a poker-playing computer program that is nearly unbeatable. What are the implications for the old debate about whether poker is a game of skill or luck
As a spectator sport, poker traditionally isn’t the most engaging thing to watch. I realised this when I found myself in the Bellagio hotel in Las Vegas a few years ago. I was there for a couple of days with a friend, and being a fan of poker, thought it would be fun to check out a high stakes game there. It wasn’t. Nothing happens – you see lots of people insist on wearing sunglasses in a windowless room, some cards and chips moving around a table, and occasionally someone being banished with a sullen look on their face, because they no longer have any chips in front of them at that particular time. Watching poker in real life is boring.
Despite that, televised poker enjoyed a huge rise in popularity in the early 2000s. Some people have argued that there are two reasons for this: table cameras, and ice hockey. In 1999, Channel 4 broadcast a programme called Late Night Pokerthat revolutionised the way in which viewers could engage with the game – by introducing under-the-table cameras so that viewers could see the cards. With this simple addition, poker suddenly became a gripping drama, where you could see the inner workings of the poker players’ minds with each and every bluff.
Then, in 2004, the North American National Hockey League went into lockout over a pay dispute between the player’s association and league officials. The cancellation of the season meant that 1,230 games weren’t played – and more critically, weren’t shown on television. Networks were left with a fairly gaping hole in their schedules that they needed to fill quickly, andaccording to Marvyn Ryder, a professor specialising in sports marketing at McMaster University, tournament poker filled that vacant space.
With the rise in televised tournaments, there was an associated rise in popularity of poker websites. And in turn, legal issues came up about whether it should be classified as a game of chance, or a game of skill. This wasn’t a new issue, as the debate about the status of poker has been going on for as long as the game itself has been around. But it’s an interesting issue – in many countries, games of chance are either illegal or regulated, whereas the rules are more relaxed for games of skill, and certainly in the case of online poker, it can be easy to abuse the system if left unchecked.
In 2012, I wrote about a study in the Journal of Gambling Studies, which argued that poker isn’t the game of skill that many players make it out to be. In the experiment, 300 participants were divided into “expert” versus “non-expert” groups, depending on whether they had an interest in the game or not. Then, they played 60 hands of Texas Hold’em in which the deals were fixed, so that players could get consistently good, bad or neutral hands. In a nutshell, the researchers found that there wasn’t much difference in the final amounts of money that the experts accrued compared with the non-experts, with the implication that skill level didn’t have much effect on the outcome. In other words, they argued, poker is a game of luck.
As I and others mentioned at the time, the paper was not without some methodological problems. One issue was that the authors seemed to miss what is perhaps the most important part of poker –the human element. I found this out first hand a few years ago, when I really got into the wholetheory behind poker, and how to figure what the best plays would be in different sorts of situations. A lot of it involves reading the other players at the table – do they play conservatively, or fairly loose? Do they “bully” the pot, or only ever make small bets? Over time, if you play with someone long enough, you can build up a profile of them, and that helps to inform when you should and should not make bets. Then, I went into an online poker room, and all of that theory got completely blown away. When you’re playing online for very small amounts of money – 1p or 2p bets – no one really cares about theory. If someone’s got what they think is a good hand, it doesn’t matter how you’ve been trying to present yourself at the table, they will just call your bets all the way.
Another problem was that poker is rarely ever played as a short-term game. Anyone who’s ever had even a semi-serious attempt at trying to make money out of it knows that if you want to make actual money playing poker, it’s not going to happen inside 60 hands. How many hands you have to play in order to become an expert is anyone’s guess, but some players report playing anything from 200 to about 10,000 hands per day. In other words, if you want to be a brilliant poker player, it’s going to take you a long time. Unless you’re a computer, in which case it will take about two months.
Last week, a study in Science reported a new computer algorithm called counterfactual regret minimisation, or CFR, that was used to weakly solve a variant of poker called heads-up limit Texas Hold’em. In other words, researchers claimed to have developed a computer program that, although it won’t win every hand it plays, it will play a near-perfect game of poker that makes it pretty much unbeatable. To do this, it spent two months playing through a billion billion hands, and built up a monstrous, 11-terabyte database of every possible combination of hands, plays and outcomes. You can try your chances against the program, called Cepheus, here.
While some people aren’t convinced that the program is truly unbeatable, the fact that researchers are able to develop an algorithm that goes some way to solving poker shows that it can’t all be down to chance. That’s not to say that luck doesn’t play a part – if it did, Cepheus would likely win every hand, every time. And while it is being lauded as a huge step in our knowledge of artificial intelligence, there are potential legal and mental health implications here too. Cepheus arguably reopens the old debate about whether the game should be classified as one of skill, or one of chance. Of course it’s both, but the question is whether it is more one or the other – and the fact that a nearly unbeatable computer program can be developed clearly shows that skill is a major factor. The worry, then, is that conclusively classifying poker as a game of skill might open the floodgates for commercial operators. Not a problem in itself, but definitely an issue when it comes to concerns about the devastating effects gambling addictioncan have.