Poker Robots?

Josh Wright —  9 May 2006

Tim Harford’s article in the Financial Times (HT: Tyler Cowen at Marginal Revolution) contains the following quote from Darse Billings on the likelihood of a poker “robot” ever beating world class poker players:

I believe that bots will eventually play better than all human beings.”

Poker superstar (and UCLA PhD) Chris Ferguson apparently agrees:

If poker robots had a tenth of the resources that were spent on chess, they’d already have beaten us.”

The article goes on to suggest that the introduction of “bots” into online play will ultimately deliver the death blow to the lucrative online poker business.  I don’t know a great deal about recent developments in this technology, but my prior is to be quite skeptical (and indeed the article suggests that, as of now, most online bots are giving away money to online players).  Along those lines, I came across the following quote from the great Doyle Brunson on the subject back in 1978 in the original Super System:

The difference between playing good poker and playing good blackjack is as vast as the difference between squad tactics and grand strategy in warfare.  You can beat a blackjack game by knowing exactly what to do in every situation … and doing it.  That’s tactics.  But in poker you may face an identical situation twice against the same opponent, handle it two different ways, and be right both times.  That’s strategy.  And that’s why there’s never going t obe a computer that will play world class poker.  It’s a people game.  

A computer could be programmed to handle the extensive mathematics of a poker game.  But the psychological complexities are another matter . . ..  A computer could play fair-to-middling poker.  But no computer could ever stand face to face with a table full of people it had never met before, and make quality, high-profit decisions based on psyschology.”

2 responses to Poker Robots?


    nesmith ankeny wrote a book on poker bluffing that allowed me to program bluff into a program for playing pot-limit five-card draw.


    I remember when IBM’s Big Blue computer first beat Gary Kasparov in chess, when he described it as ‘like playing God’ and chess experts were amazed to see Big Blue could continue to make creative moves with very few pieces on the board, when chessmasters would universally say so few pieces left the game inherently unwinnable. But Big Blue could foresee so many more possible tactical moves than a human, that at that level, *tactics became strategy.* And hence Big Blue won.

    To my view, poker is very different because it is strategy from the start. How do you teach a computer to bluff? How can a machine learn the difference between an apparently stupid move from a beginning player that actually *is* stupid, and apparently stupid move from an expert that actually is brilliant? So I’m inclined to agree, that computers playing poker as well as they play chess is a long, long way off.