I can still remember back to 1996 when former World Chess Champion Gary Kasparov took on an IBM computer by the name of Deep Blue in a famed match that made news headlines everywhere. Kasparov won the famous chess match 4 games to 2 supposedly proving that no computer could beat a human at the game.
Unfortunately, the Russian agreed to take IBM up on its offer again in playing the machine a second time. The upgraded version of Deep Blue (nicknamed Deeper Blue) got revenge for the previous loss and won two games to Kasparov’s one with three games being draws.
What’s this have to do with poker? Well for the second straight year, the Polaris poker playing machine will be brought out to compete against some of the best poker players in the world. Last year saw Phil “The Unabomber” Laak and Ali Eslami defeat the 1st version of Polaris two games to one with one game being a tie.
Now, however, the machine is upgraded and comes with a new name: Polaris 2 (flashbacks to Kasparov and the chess matches?). And this time the opponents for the computer have changed too as Polaris will be facing off against some of the best online players in the world.
The online players definitely know what they’re up against too as Bryce Paradis, who cashed for over 3 million dollars last year, had this to say, “Against the current AI in Polaris 2 the average poker player would be completely dominated. The Polaris 2 team has made incredible improvements since the match last year.”
Let’s hope that the professional poker players such as Paradis won’t be completely dominated by the machine. Otherwise, they’ll suffer the same humiliation that Kasparov suffered back in 1997 when he became the first World Champion to lose to a machine.
I hope the rise of the poker machines will be adverted by the situational nature of poker. Unlike chess, there isn’t one best move to have in the database. I think with aspects like bluffing involved, computers will never get the pyscology of the game.
Hi,
My name is Mike Johanson – I’m one of the programmers on the Polaris project. I found your blog while I was procrastinating this morning before starting on the day’s bot-coding :^D
I think the attitudes of the competitors play a large role in determining how everyone looks back on the event. The second match between Kasparov and Deep Blue didn’t end well – Kasparov was convinced the IBM team had cheated by having humans intervene during the match. He said Deep Blue had made “human” moves, that only a human (and never a computer) would make. There just isn’t any such thing – and likewise, Grundy is wrong about the psychology of bluffing being a human phenomenon. Bluffing is actually the easiest part for us, because the mathematics of game theory tells you how often to bluff and slowplay.
So, Kasparov insisted that IBM had cheated, and whether it was true or not (and I’m sure it isn’t, but hey – I’m on the computer’s side, right?), it looked like sour grapes. If he was “humiliated”, then that was a big part of it.
Poker has been different so far. The work we’ve done in poker over the last 14 years has turned up a lot of knowledge about the game that humans are able to use to improve their game. I think the pros we’ve played against (Laak, Eslami, Paradis) have learned just as much from the matches as we have. If this improves their game when they play against other humans, then I think it’s a win-win situation.
We’re certainly not out to try to humiliate anyone. By doing this research, we’re learning things about AI that apply outside of Poker and outside of games. The friendly competition we get from playing against poker pros is a) really motivating, b) good for research, and c) awesome and fun. I think the pros feel the same way.
Hi Mike,
You make a very good point about the attitudes of the competitors playing a big role in how the event will be perceived later on. Kasparov didn’t have the best attitude after losing and I agree that it’s doubtful IBM had humans intervene (and if they did shouldn’t Kasparov been able to beat them since he was a World Champ?).
By saying “humiliated” I meant that, instead of being recognized as one of the greatest chess champions of all time, he was instead presented through the mainstream media as the chess champ who lost to a computer.
Maybe if Kasparov would have been more gracious in defeat, he wouldn’t have been shown in that light. By accusing IBM of cheating, Kasparov may have brought on plenty more negative attention.
If your poker machine does beat Paradis and the rest of the online pros, hopefully they’ll take defeat a lot better. From the quotes I’ve read by Paradis, he seems to be taking this competition in a friendly manner.
And about Grundy’s comment, I’m sure you know plenty about what Polaris is capable of since you’re one of the programmers and I believe what you say about the bluffing aspect. But, before reading your comments, I would have been in the same boat as Grundy in thinking that a computer wouldn’t be able to handle situational play such as bluffing, slowplaying, and profiling opponents.
Anyways, good luck with the competition and I look forward to seeing how it turns out.