Beginning on the 11th of January 2017, four of the world’s top poker pros will take on an Artificial Intelligence program in Heads-Up No-Limit Texas Hold’em. The event will span over 20 days and consists of a collective 120 000 hands. The four professional poker players chosen to compete against the AI for a share of $200 000 are Jason Les, Dong Kim, Daniel McAulay, and Jimmy Chou.
The AI program is named Libratus and was developed by a team at Carnegie Mellon University. However, the Libratus is not the first AI of its kind. Its predecessor is Claudico. Claudico took on four prominent professional poker players in 2015 when 80 000 hands were played in Heads-Up No-Limit Texas Hold’em. It was then decided that 80 000 hands were too few to officially assess who would be the better player, man or machine.
The AI’s Strategy
While games like chess may be easy for computer programs to win, there is quite a difference between chess and poker for an AI program. The difference lies in the fact that chess is a game of complete information and poker is one of incomplete information. A game of complete information means that each player is aware of all the moves that have been made and is also aware of what available moves are left. A game of imperfect-information (like poker) means that not all the information and moves that have been made are available to all the players in the game. In a game like chess, the AI can strategize several moves in advance whereas a game of imperfect-information makes it challenging even for an AI program.
Instead, the AI will play in a way that isn’t expected by humans in a game of poker. It will make moves that poker pros around the world would consider as bad ones. This may spur human players to adopt similar unconventional strategies like the AI program.
The ultimate test between man and machine will lie in the results of the event. We all wait with baited breath as the poker pros take on Libratus this month.