It used to be that computer programmers were trying to develop the perfect computer chess game, which they did in 1997, a program that managed to beat some of the world’s top chess players.  Today, they have turned their attention to poker, namely the Texas Hold’Em variety, and they are trying to develop the poker program that will be able to account for pocket aces, open-ended straight draws, flops, rivers, and anything else that goes into a hand of poker.

They’ve come pretty close, producing an AI machine that was able to beat some of the top online poker players in a contest known as ‘Man vs. Machine’.  This second generation machine showed up at the World Series of Poker in Las Vegas and proved it had the right stuff.  Another program, called ‘Fell Omen 2’, took second place in an annual contest that pits AI poker programs against not humans, but one another.  Developed by Ian Fellows, a researcher and statistician at UC San Diego Department of Psychiatry, the programs supposedly ‘has no weaknesses’.  It is free for poker players to use on Fellow’s web site.

What is interesting to note about these AI poker programs is that they can determine the angles and odds of the game, bluff and bet accordingly.  That is all well and good, but the programs do not get all of the information they need and they cannot account for the ‘human factor’ that is very prevalent in the game.  While the game of chess only has so many set moves, the game poker has more variables that require the quicker thinking of a human mind.  Remember, a computer can only ‘think’ as fast as its processor will allow.

Additionally, computer programmers do not have the advantage of being able to look at their opponents and read their body language or hear the tones of the voice.  The programs are based on mathematical calculations, one-on-one play and simple betting structures.  Unfortunately, real life play is very seldom that straight-forward, not when you have players that hide their eyes by wearing sunglasses and hoodies at the poker tables.  A good poker player will observe their opponents and watch what they do to make their decisions.  A computer program can only go based on the cards it ‘holds’.

While the idea of an AI poker player is good and will challenge players who use it to practice, the fact that they win as often as they do may have more to do with lady luck being on its side instead of the human player’s side.  Of course, the idea behind the computer poker players is not just for beating poker players.  The information the researchers gather while working with the AI intelligence will go a long way towards other applications.