Last year's match

Last year the predecessor of this match took place, which was decided in favor of the humans that time, being Phil Laak and Ali Eslami. Since then the programmers of the University of Alberta have been making radical and innovative improvements to the robot.

Limited Information

A robot that manages to take on a team of professional players is shocking news for the poker world, and quite a triumph for the scientific community. Checkers and chess players already knew they were no match for their electronic counterparts, but poker seemed a whole different game since it is a game of limited information. In chess and checkers both players can see the board, yet in poker you only have access to your own cards.

Reaction of the Researchers

Michael Bowling from the University of Alberta comments: "We are very excited. I think there's no question that we're playing on the level of the absolute best human players in the world at the game of heads-up limit hold'em"

Bowling furthermore declared that improvements will have to be made especially concerning the computer's capabilities to learn the opponent's strategy and to adapt to it. Nevertheless, Bowling called this achievement of Polaris a 'crossing point' in the poker world, since now computer programs are to be considered equal to the best human players.

Specific Poker Variant

The poker game that was played is Limit Texas Hold'em. This is a very specific form of poker, perhaps the most basic form for a computerized robot to analyze and strategize against. Limit Hold'em has fewer variables to be calculated than No-Limit Hold'em, simply because the amount to be bet is fixed.

Specific Type of Play

The fact that the heads-up play is limited to only two players also adds to its simplicity. Putting in one extra player would make the game many times harder to calculate for any artificial intelligence program, since then far more combinations would be possible. Let alone when a game is played with 10 players, as is often the case with poker tournaments and cash games. A robot that can manage to survive in such a game will take years and years to be developed, if it can be developed at all.

Limitations

It is certainly true that this is an amazing achievement of the team of scientists that created this artificial intelligence program. Only, to say computer programs are to be considered equal to the best professional players goes a bit far. Bowling has to keep in mind that the Polaris project is limited to a very specific variant of poker, and to an even more restricted type of play, namely Heads-up.

Scientific Bragging

Limit Heads-up play is one of the simplest forms of poker from a game theorist point of view, what is off course the reason the scientists started with it. Nothing wrong with that, however stating that 'computer programs are to be considered equal to the best human players' to me sounds a lot more like bragging than the deduction of a sound scientific conclusion.