Overreacting to a Computer Beating Poker Pros

Newspapers trumpeted a landmark event last week: a computer program beating professional poker players head-to-head at Limit Hold-Em. Parallels have been drawn to Big Blue‘s victory over Gary Kasparov roughly a decade ago.

Those parallels are not very meaningful.

First, heads-up Limit Hold-Em is a very simple version of poker — exactly the kind of game that a computer should be good at.

Second, unlike chess, which is deterministic, there is an enormous short-run luck component in poker. The variances are so high in poker that you need weeks if not months of play to confidently determine whether you are a winning player. There is no way with standard poker and two competent players that you could determine who is better in a short period of time. To try to remedy this shortcoming, the man vs. computer contest was actually run differently to lower the variance according to this report:

The matches will be duplicate matches of 500 hands of heads-up Limit Hold-Em. Each match will involve two separate human opponents playing simultaneously against Polaris. The same hands (hole cards and board cards) will be used in both matches but the humans will be dealt opposite sides of the cards. So if on the first hand one of the humans is dealt aces on the button, then Polaris will be dealt aces on the button in the first hand against the other human. If one player gets lucky in the 500 hands, their partner necessarily will be getting bad luck in their 500 hands, helping reduce to the luck factor in deciding a winner. The winner is determined by adding together the teams’ winnings (or losses). If either side [wins] by 25 small bets or more in total, then they are the winner, otherwise it is considered a draw.

This is a very clever structure (based on duplicate bridge) and greatly reduces the variance by making sure that the pairs of humans and pairs of computers receive the same set of cards. I have never seen any calculations done, but I suspect that even playing poker with duplicate hands has a huge variance in outcomes over such a short time horizon.

I actually think there are two things coming out of this poker contest that are more interesting than who won or lost: the design of the contest is very interesting, for instance.

Whenever someone wants to learn “the truth” in a world beset with randomness, reducing the variance is a critical part of solving the problem. This structure of poker where the luck of the cards is eliminated is a great example.

I could imagine lots of different settings in business or life where one could steal the insights of this duplicate poker design to learn more quickly about the world.

The second thing I find really interesting is the following: If duplicate poker reduces the luck factor in poker, is there a broader market for duplicate poker? You might think so. There is a website online where you can play duplicate poker, but I’ve never actually heard of anyone playing there.

Why is that?


Mike B

One of the best things a computer can bring to the poker table is the ability to inject randomness into the play. Typically, poker seems to be less about the cards and making moves based on probability, but being able to figure out what your opponent is thinking. By being able to smartly inject randomness into decisions, what I like to call the Ronald Regan strategy, you can completely throw your opponents for a loop and cause them to self-destruct. The trick is using just enough randomness so that you avoid just making stupid moves and self destructing yourself. Once someone is able to model that, computers should be able to beat humans fairly consistently.

What is perhaps a much more interesting topic for this website is studies on collusive playing, which is now much easier with the rise of non-face to face internet poker. Questions would include how the advantage changes with the number of confederates, the optimal strategy they should employ and the best way to avoid detection.

Read more...

ian

@AaronS: Yes, the computer can figure the odds more perfectly, but since there is betting involved in poker, it doesn't matter as much. The computer might win more hands, but unless it bets properly, it will still lose. Note that this isn't *that* much of an advantage, either; professional poker players can calculate odds quite accurately, as well, even if they can't quite reach the precision of a computer.

And, obviously, the computer can't just bet according to the odds. Such behavior would give a tremendous amount of information to a human player, who could exploit it. The program must be more subtle than that.

Ryan

What stops them from doing a chinese room?

Wait to see what Player A does against computer, then make that move against player B.

frankenduf

duplicate poker would be boring- it's precisely the luck factor that makes poker fascinating- how the pros weather the storms lady luck throws down is the measure of their determination- duplicate backgammon would be boring for the same reason

AaronS

Isn't poker, to a degree, somewhat deterministic, as well? I mean, each hand has certain odds--odds that a computer could figure to near perfection.

The computer knows what cards it has.

The computer knows the odd, after each turn of the card, as to having a winning hand, it would seem.

It's not a science, but as to probabilities, couldn't the computer make wiser decisions than a human--and CERTAINLY less emotional ones? While the human might be faked out by cool playing, the computer if basing everything on probabilities...and we would expect it to win, no?

Vinod

FTA:
"Human pride took a hit 11 years ago when IBM's Big Blue computer beat world chess champion Gary Kasparov"
Actually I am proud that we were able to build a computer that can beat the best of humans in chess. That is obviously progress unlike relying on nature to toss up a few Kasparovs and Anands every now and then.

mikelotus

And the luck in poker is the exact reason I no longer play chess since it is deterministic. I play strategy board games, usually of the military type, but there are dice, cards, etc., that add luck, but the luck is actually realism. The fog of war, unexpected events are like real like and how you adjust on the fly and think your opponent will adjust on the fly is a key component. Many Europeans prefer the games where there is no luck what so ever, but once again -- boring. And until there is artificial intelligence, we won't see a computer that can beat a competent player in a true strategy game like I am referring to because, well I don't know why, but Roger Penrose has speculated because that type of thinking happens on the quantum level.

Tom

The main problem with this test of the computer's ability to beat the pros is that limit hold'em is rarely played heads up. Typically, the game is played with eight to ten players; therefore, any meaningful test of the computer should incorporate more human players at the table. I suspect the computer would almost always determine it was beat due to the mathematical odds of having to beat so many players. If my guess is right, the computer would fold too often, and the human competitors would win a large majority of the time.

www.PlayingCardsAndMore.com

anmol

Collusion is possible while playing duplicate poker, u can call up your partner and tell him the other guys hole cards.

Tommy

The whole of playing cards is to spend time with friends and family, why play a computer?

http://thebridgesource.com