Bookmaker News

Two computer specialists, Evan Hurwitz and Tshilidzi Marwala, who are researchers at Witwatersrand University in South Africa, have hinted that they are on the verge of a breakthrough regarding the ability of computers to bluff at poker.

Poker lovers are intrigued to find out whether machines with artificial intelligence can bluff and even play against all the odds, especially given that these “robots” have a definite advantage: they show no signs of nerves and have no facial expressions to betray their anxiety.

The two researchers have shown that, in addition to being capable of bluffing, computers can even develop their own strategies!

The researchers decided to match a poker agent and two robots against two other similarly trained robots in order to see what would happen, without them knowing the rules of the game.

The robots gradually started to develop their own poker-playing personalities, which were aggressive or conservative according to their past performances. The most remarkable of these robot players, named Randy, even managed to bluff: “Although he had one of the weakest hands, he chose to play on against all the odds which were telling him to fold”.

In the same vein, a similar type of challenge will take place in Vancouver, on July 23rd and 24th 2007. This will involve two poker professionals and bluffing experts - Phil Laak and Ali Eslami, going head-to-head against Polaris, another machine with artificial intelligence, created by developers at the University of Alberta!