Online Poker Played by Bots

I forwarded a recent slashdot link to an LA Times story on online poker bots to my cousin Peter Balka, who earned his way to his PhD by playing poker. The slashdot article claims, “many engineers also are trying to make bots that are good enough to play and beat human players for money in online casinos.” Peter replied:

I strongly suspect that there are already more than a handful of
first-generation bots currently playing internet Hold’em poker. The game
is simple enough (and the players bad enough) that even the most
rudimentary bots would undoubtably churn out a considerable profit by
playing only premium starting hands on multiple tables at multiple sites
24/7.
 

Darse Billings, a student at the University of Alberta, has centred
his PhD thesis on artificial intelligence on a poker Deep Blue he
developed that he claims has been successful making money on the internet.
[A copy of a National Post article on Billings’ work is available from the University of Alberta website.]
I have no reason to doubt his claim.

The major difference between Deep
Blue and any poker equivalent is that, minor fluctations in day-to-day
mental processing aside, chance or ‘luck’ does not play a significant role
in chess, whereas it is the overriding short term factor in poker. For
this reason, any results from a bot vs. bot or bot vs. professional
tournaments I would certainly take with a grain of salt, just as I do the
crowning of a new ‘world champion’ each year at the World Series of Poker.
I would be more comfortable in drawing any kind of conclusions from such a
tournament if the participants were to play the equivalent of 100,000
hands or so, not the 100 – 200 that they would typically see in the course
of a tournament. I would equate any conclusions as to the best poker bot
based on the results from this Poker Bot Tournament to my obtaining 3
heads in 5 tosses of a coin and concluding that my coin is biased.