News

[November 2010] With the World Series of Poker final table just getting underway, The Register has published an article about the CPRG and how far away computer bots are from being able to compete in such an event.
[October 2010] Peter Brown from CBC Radio recently interviewed several members of the CPRG for the "Ideas" program. The interviews are compiled into a show called Microchips that investigates poker strategies, Polaris, and how they relate to everyday decisions.
[July 2010] The 2010 Computer Poker Competition has come to an end. This year's competition used a similar format to last year's, with 2 different divisions in each of the 3 different games played. There were several entries again from both researchers and hobbyists. The CPRG picked up three 1st place finishes to go along with one 2nd place and two 3rd place finishes. The top 3 programs in each of the 6 divisions were as follows:
Heads-Up Limit Texas Hold'em:
Total Bankroll
1st: PULPO (Marv Andersen, UK)
2nd: Hyperborean-TBR (CPRG)
3rd: Sartre (University of Auckland, New Zealand)

Bankroll Instant Run-off
1st: Rockhopper (David Lin, USA)
2nd: GGValuta (Mihai Ciucu, Romania)
3rd: Hyperborean-IRO (CPRG)
Heads-Up No-Limit Texas Hold'em:
Total Bankroll
1st: Tartanian4-TBR (Carnegie Mellon University, USA)
2nd: PokerBotSLO (Universities of Maribor & Ljubljana, Slovenia)
3rd: HyperboreanNL-TBR (CPRG)

Bankroll Instant Run-off
1st: HyperboreanNL-IRO (CPRG)
2nd: SartreNL (University of Auckland, New Zealand)
3rd: Tartanian4-IRO (Carnegie Mellon University, USA)
3-Player Limit Texas Hold'em:
Total Bankroll
1st: Hyperborean3P-TBR (CPRG)
2nd: LittleRock (Rod Byrnes, Australia)
3rd: Bender (Technical University Darmstadt, Germany)

Bankroll Instant Run-off
1st: Hyperborean3P-IRO (CPRG)
2nd: dcu3pl-IRO (Dublin City University, Ireland)
3rd: LittleRock (Rod Byrnes, Australia)
For complete results of the competition, see the competition results page.
[May 2010] The CPRG has released a new paper: Using Counterfactual Regret Minimization to Create Competitive Multiplayer Poker Agents, that appeared at AAMAS 2010 in May. The paper demonstrates that counterfactual regret minimization (CFR) computes strong poker strategies in 3-player games, even though game-theoretic guarantees are lost outside of the two-player case.