The First Man-Machine
Poker Championship

The Competitors

The Humans

Phil Laak

Phil Laak is one of the most recognized professional poker players in the game today. His trade-mark grey hoody and sunglasses gave him the nickname "The Unabomber". Phil's enthusiasm, energy, and playful antics have made him a crowd favourite everywhere he goes.

Phil's charisma also makes him a great showman. When he played against the U of A program Poki-X in Las Vegas during the 2005 World Series of Poker, the audience was enthralled and amused by Phil's running commentary, which was both informative and entertaining. Recently he has applied that talent as the host of a celebrity poker television show.

But behind the flamboyant image is a cool calculating engine. Phil earned a degree in mechanical engineering before becoming a top-flight backgammon player -- a game that is much more about probabilities than psychology. His analytical skills were amply demonstrated in the Vegas'05 poker match. He took that match very seriously, training against the AI players in Poker Academy for two days prior to the match, probing for any weaknesses that might be exploited. We have no doubt that he will be fit and ready to play against the new generation of bots in July.

Ali Eslami

When Phil was asked to select a partner for the duplicate matches, his immediate first choice was Ali Eslami. Why did he choose a relative unknown when Phil is so well-connected to many of the elite players in the poker world? Simply because he believed Ali was the strongest teammate he could have for this competition.

Ali is an extremely successful high-stakes cash game player, playing in some of the biggest games in the world, from $400-800 betting limits up to $1000-2000 per bet. He plays a wide variety of games, but is particularly adept at short-handed and heads-up Limit Texas Hold'em. Although he travels the globe playing in ring games and the occasional tournament, Ali's "home office" is the largest poker room in the world, the Commerce Casino in Los Angeles, California. He has recently acted as guest commentator for a popular television poker series, sharing his profound understanding of poker strategy and tactics.

Ali is a sterling example of the new generation of sharp, young, technically oriented, and mathematically savvy poker players. He worked as a computer consultant before discovering his prodigious talent for poker. Ali is very familiar with the state-of-the-art in poker bot technology, having read (and understood!) the CPRG technical papers. That means he knows their strengths and weaknesses, and can exploit their vulnerabilities. He is also a deep strategic thinker, and can accurately predict the match strategies the other side might employ -- even before they've thought of them!

The Humans as a Team

Phil Laak and Jennifer Tilly
Phil battles the bots in Vegas'05, while Jennifer Tilly looks on.

How do we think Phil and Ali stack up as a team? In many ways, these two players are our worst nightmare. They are super strong players who understand the mathematical foundations of the game (as opposed to players who specialize in understanding the particular foibles of human opponents). That means they can cope with anything a bot might throw at them, whereas many of the "old-school" players would be ill-equipped to handle some of the inhuman (but perfectly valid) approaches to the game that bots are capable of employing.

Frankly, we would like our chances a lot better against any two of the game's legendary players. We might get more notoriety by doing well against, say, Doyle and TJ, but as great as those players are, we honestly believe they wouldn't be as hard to beat as Phil and Ali.

And that's just fine with us, because one of the ultimate goals of the research is to develop a program that plays better than all human players. We're delighted to have the opportunity to play against some of the toughest opposition that Carbon can muster against Silicon.

The Computer Team

University of Alberta Computer Poker  Research Group Photo
Members of the University of Alberta Computer Poker Research Group
Standing: Mike Johanson, Josh Davidson, Marty Zinkevich, Martha Lednicky, Neil Burch, Robert Holte, Jonathan Schaeffer, Michael Bowling, John Hawkin.
Sitting: Nolan Bard, Morgan Kan, Darse Billings.
Missing in Photo: Andrew Albert


The University of Alberta Computer Poker Research Group (CPRG) has been recognized as the world leader in poker-playing programs for more than fifteen years.

Their research has won awards and acclaim from the Artificial Intelligence community for significant contributions to the science of information. They have been encouraging (and winning) friendly competitions between 'bots' since 1998. Some of the past CPRG bots are available to everyone, inside the best poker training software on the market, Poker Academy.

In July of 2006, the Alberta team won the world championship for computer poker programs, held at the annual conference of the American Association for Artificial Intelligence. In two tournaments that took weeks of continuous play to complete, the U of A won every match they played by a large margin, thereby adding another title to their long-standing domination.

Their previous programs have already proven to be a worthy challenge for elite players. In 2003, top online pro (and short-handed specialist) Gautam Rao (aka "thecount") played a 7000-hand match against PsOpti-1, but was not able to win by a statistically significant margin.

In 2005, Phil "The Unabomber" Laak played a very short match against Poki-X (an experimental mixture of Sparbot and Vexbot), winning the coin-toss, but not before having his nose bloodied. In one of the match highlights, facing a check-raise on the river, Phil thought for a long time before proclaiming: "If that is a bluff, it's over for humanity". He folded. But it was a stone-cold bluff, from a stone-cold poker face that no human could possibly read.

Polaris is the latest generation in a long line of programs stemming from the ongoing CPRG research. The time has come to gather some more scientific data, in the most serious match to date.