How Computer Programs Modeled Poker


Poker is a card game in which players place bets on the strength of their hands and other information about the others at the table. The game requires a strong understanding of probability, psychology and game theory. Players must also have good emotional control to avoid becoming frustrated and making bad decisions.

Each player begins a hand by placing a bet, usually in the form of chips representing money, in the pot. A token called a button (or buck) rotates clockwise among the players to indicate a nominal dealer, who handles the cards for each round of betting. Once each player has a complete hand, the cards are revealed and any one player who makes a best five-card hand wins the pot.

When it’s a player’s turn to make a bet, they can say “call” to match the amount of the last bet or raise by a certain amount (usually two chips before the draw and four chips after). Players can also fold their cards, forfeiting any chance to win the hand.

The success of von Neumann’s proof paved the way for a future in which all sorts of competitive interactions could be mathematically modeled and analyzed—including auctions, submarine warfare and even the ways species compete to pass their genes on to future generations. However, in the field of games research, poker made little headway until a team of researchers from the Department of Computer Science at the University of Alberta began work on a program known as PioSOLVER in the early 1990s.