Bot Beats Human Champs at Poker
Las Vegas (NV) - An unflinching, uncaring silicon machine beat several top poker champions during "Man versus Machine" match at the Rio Hotel in Las Vegas.
The Polaris program developed by University of Alberta researchers was beaten last year, but managed to best the players this year with a 2 win, 1 loss and 1 draw record. And don’t think the machine had it easy, three of the human competitors had more than $1 million in lifetime winnings each.
Each of the four matches consisted of 500 hands of Limit Texas Hold’em poker. Researchers analyzed approximately eight billion poker hands and believe the program can play a perfect game.
"It’s possible, given enough computing power, for computers to play ’perfectly,’ where over a long enough match, the program cannot lose money," said Michael Bowling of the university’s computer poker research group. "Humans will always make some mistakes, meaning the program will have an advantage."
-
Previous News Article
$1000 Belkin Box Promises... -
Next News Article
Microsoft to Cut Xbox 360 Price...
Of course! The Bot can count cards perfectly.
Of course! The Bot can count cards perfectly.
Ummm, counting cards is for blackjack. When you reshuffle after each hand, then counting doesn't apply.
It still lost rounds so its still only as good as being lucky.
The thing is though it was playing limit texas hold'em which is limited to just strict odds only. There is not the finesse you get with no limit where you can have a losing hand and out bluff someone to win the pot. With a pot limit texas hold'em best had wins period not the best "played" hand
Keep in mind that this bot was programmed for limit Hold 'em, which relies more on probability than no-limit. Also, even playing perfectly you will still lose at least 20% of the time.
Try programming the bot for no-limit and professional players will still win.
This only shows that playing proper limit hold 'em will prove successful, but this does not mean that the bot is playing entirely optimal.
Keep in mind that this bot was programmed for limit Hold 'em, which relies more on probability than no-limit. Also, even playing perfectly you will still lose at least 20% of the time.
Try programming the bot for no-limit and professional players will still win.
This only shows that playing proper limit hold 'em will prove successful, but this does not mean that the bot is playing entirely optimal.
A couple of other comments, the matches were played heads up (1 on 1) instead of a more traditional of 6-9 players and a small sample of only 500 hands.
Also, there are bots out there that play no limit but they generally employ a fairly basic strategy which is easy to exploit once its identified.
Well, I play $6/12 limit a LOT locally. Try putting this bot on a 10 seat table with a mix of players, including the total donkeys that play every hand and chase gutshots and other long shots to the river. See what happens when pocket aces have a 25% chance to win.
I have found that there is no strategy that will work when you can't hit your draws (like flushes and open ended straights) and others are hitting two outers on you.
I would like to see if this bot can prove the concept stated in books that you will win in the long run if you play right. Can it come out ahead given the usual rake a card room takes per hand along with the fluctuations of "luck" that occur?