|
|
Poker Related Links
|
|
Individual Poker Pages
rec.gambling.poker (Usenet)
Commercial Services
Google Index
|
|
|
Recent News
|
- [December 2008] The December 2008 issue of WIRED magazine has a story about Polaris' 2008 match
against the Stoxpoker pros, told in comic book format: Ante Up, Human:
The Adventures of Polaris, the Poker-Playing Robot
- [September 2008] With the summer competitions finished, the CPRG has been writing papers to
present the research that created Polaris. The most recent paper,
"Strategy Evaluation in Extensive Games with Importance Sampling" was presented at the ICML 2008 conference in
Helsinki, Finland the day after Polaris' victory at the Second Man-Machine Poker Championship. This and two
other recent papers related to Polaris, "Regret Minimization in Games
with Incomplete Information" and "Computing Robust
Counter-Strategies", have also been added to the Publications section.

| Match Number
| Player |
Amount Won |
Player |
Amount Won |
Difference |
Result |
| Remote 1 |
Matt Hawrilenko |
+$199500 |
IJay Palansky |
-$174000 |
+$25500 |
Humans Win |
| Remote 2 |
Nick Grudzien |
-$2000 |
Kyle Hendon |
-$118000 |
-$120000 |
Polaris Wins |
| Live 1 |
Nick Grudzien |
-$42000 |
Kyle Hendon |
+$37000 |
-$5000 |
Draw |
| Live 2 |
Rich McRoberts |
+$89500 |
Victor Acosta |
-$39500 |
+$50000 |
Humans Win |
| Live 3 |
Mark Newhouse |
+$251500 |
IJay Palansky |
-$307500 |
-$56000 |
Polaris Wins |
| Live 4 |
Matt Hawrilenko |
-$60500 |
IJay Palansky |
-$29000 |
-$89500 |
Polaris Wins |
- [July 2008] In dramatic fashion, Polaris wins the 2nd man-machine
showdown with a record of 3-2-1. Polaris played against some of the
best heads-up limit players in the world and came out on top with a
decisive win against Matt "Hoss_TBF" Hawrilenko (left) and IJay "doughnutz"
Palansky (right) in the final match. For more information on the match, please
visit our man-machine
page.
- [June 2008] The CPRG is proud to announce that they are once again
taking on some of the world's best poker players in a man-machine
showdown. In a partnership with www.stoxpoker.com, the CPRG will face
off against some of the world's best online pros. Nick "stoxtrader"
Grudzien, Matt "Hoss_TBF" Hawrilenko, and IJay "doughnutz" Palansky are so far slated to play against Polaris
at the 2008
Gaming Life Expo in Las Vegas, July 3-6. More details for the match
can be found on stoxpoker's man
machine webpage, and our own man-machine webpage.
- [March 2008] A webapp version of the
DIVAT tool is now in beta testing.
The DIVAT tool provides an unbiased, lower variance estimator for
money. This allows you to determine the winner of a poker match with a
smaller sample size or more accurately. For more information on how the
tool works, please consult Morgan's M.Sc.
thesis or Darse's Ph.D.
thesis.
- [March 2008] Nolan's M.Sc.
thesis has been added to the publications section. It investigates
several techniques for opponent modelling in Kuhn poker.
- [October 2007] Michael's M.Sc.
thesis has been added to the Publications section. It presents some
of the techniques used to build the Hyperborean (AAAI 2007) and Polaris
(Man-machine match) bots.
- [July 2007] The first
Man vs Machine Poker Championship was played this summer.
Polaris gave poker pros Phil Laak and Ali Eslami
a tough time, but the pros squeaked out a 2-1 win with one draw. For detailed
match and post-game coverage visit the
Match Results Page.
| $10/$20 Limit Texas Hold'em |
Session 1 |
Session 2 |
Session 3 |
Session 4 |
Sum |
| Ali Eslami |
+$395 |
-$2495 |
-$635 |
+$460 |
-$2275 |
| Phil Laak |
-$465 |
+$1570 |
+$1455 |
+$110 |
+$2670 |
| Combined Score |
-$70 |
-$925 |
+$820 |
+$570 |
+$395 |
| Result |
Statistical Draw |
Polaris Wins |
Humans Win |
Humans Win |
Humans Win |
- [July 2007] The University of Alberta team won two of three
events at the second
AAAI poker bot competition. The University of Alberta team
took first place in the
equilibrium competition, first and second in the
bankroll competition, and third place in the
no-limit competition. The competition featured 43 bots from 15 teams.
Limit Equilibrium:
1st: Hyperborean07 (U of A)
2nd: Ianbot (Ian Fellows, UCSD)
3rd: GS3 (CMU)
Limit Bankroll:
1st: Qeqbot (U of A)
2nd: Hyperborean07 (U of A)
3rd: GS3 (CMU)
No Limit Equilibrium:
1st: BluffBot 2.0 (Teppo Salonen)
2nd: GS3 (CMU)
3rd: Hyperborean07 (U of A)
For more details, please visit the
tournament website.
- [June 2007] Our friends at
Poker Academy recently released a new poker analysis tool called
Poker Academy
Prospector. This software allows you to import hand histories from
many poker sites and analyze your game. A must for any serious
poker player!
- [June 2007] The first ever
Man vs Machine Poker Championship will be played this summer,
at the AAAI'07 conference. Elite pros Phil Laak and Ali
Eslami will play a series of duplicate matches against the newest
U of A program, Polaris. See the
match website for full details!
- [January 2007]
Morgan's M.Sc. thesis has been added
to the Publications section. It provides a better description of the
DIVAT analysis tool, including a simplified proof of correctness, and
applications to the Vegas'05 and AAAI'06 matches.
- [September 2006]
Darse's Ph.D. dissertation, and the
recent journal paper on the DIVAT analysis tool, have now been added to the Publications section.
- [August 2006] There was another barrage of media attention
following the UofA victory in the AAAI'06 competition. Some
local TV news
coverage is available on
youtube. Watch it and learn how to find the tumor in your
opponent's brain! :-)
- [July 2006] The University of Alberta team has won the first
ever
AAAI poker bot competition, in convincing fashion, winning
every match it played in both tournaments. The following summary
crosstables give the average win rates, in millibets (small bets
per game times one thousand):
| Normal Pace |
Alberta |
Bluffbot |
Monash |
Teddy |
Average |
| Alberta |
x |
+50 |
+720 |
+410 |
+390 |
| Bluffbot |
-50 |
x |
+530 |
-190 |
+100 |
| Monash |
-720 |
-530 |
x |
+1170 |
-30 |
| Teddy |
-410 |
+190 |
-1170 |
x |
-460 |
| Slower Pace |
Alberta |
Bluffbot |
C M U |
Monash |
Average |
| Alberta |
x |
+110 |
+180 |
+730 |
+340 |
| Bluffbot |
-110 |
x |
+120 |
+520 |
+180 |
| C M U |
-180 |
-120 |
x |
+650 |
+120 |
| Monash |
-730 |
-520 |
-650 |
x |
-640 |
more detailed crosstables (with standard deviations)
Numerous duplicate matches were played in each case, in order to obtain
statistically significant results. Due to the short timeframe we had
to prepare for the competition, we simply used a combination of two
Sparbot programs (PsOpti4 and the new
PsOpti6). The second place finisher was
Teppo Salonen, a hobbyist
programmer who built his own version of Sparbot for the
Poker Academy software platform.
A highly-publicized entry from Carnegie Mellon University did not
perform up to expectations, losing to the U of A and to Bluffbot.
No doubt the talented researchers at CMU will be back with a strong
contender next year. Their new program will also need to play
substantially faster to maintain the normal pace of Limit Hold'em.
The special slow event (one minute per game) that was created to
accomodate the CMU entry may not be repeated in future tournaments.
More details of the event are available in this
press release, and at the
competition website.
A full analysis of all matches will be performed over the coming weeks.
- [June 2006] The new poker entertainment game
STACKED, has finally been released, featuring the
Poki AI and Xenbot from
Poker Academy. The game has not been getting great
reviews, but the AI is receiving kudos from the new audience.
- [March 2006] The Daily
Planet, a science and technology news program on the
Discovery Channel, did another segment on the U of A
computer poker research group, as part of a special series on
the science of gambling. A
video clip of the segment and a
cover story are available on their website.
- [Feb 2006] A new version of
Poker Academy, PA Pro 2.5, has been
released. The best training software for Texas Hold'em just keeps
getting better. The CPRG is very proud of what Aaron, Terry, and the
rest of the PA development team have accomplished, and we strongly
recommend Poker Academy for any serious poker player.
You can download Poker Academy
and try the free demo yourself, or
click here to view screenshots of the many features.
|
|
Previous News
|
- [July 2005] A new version of the
Poker Academy software has been released.
PA Pro 2.0 is a vast improvement on what was already the
best software learning tool for the game of Texas Hold'em. The program
features our world-leading artificial intelligence (including the
Sparbot and Vexbot experimental AIs), along
with a myriad of features for poker analysis. The new program adds
No-Limit Hold'em, multi-table tournaments, and a free online poker
server. You can
download the free demo software here.
[view screenshots]
-
[July 2005] World-class expert Phil "The Unabomber" Laak
played an exciting match against our experimental Poki-X
AI in Las Vegas (while the main event of the World Series of Poker
was taking place just around the corner). Unfortunately, a short
match with rapidily escalating blinds makes the outcome little more
than a random coin-flip. The full story and a detailed analysis of
the match will be made available in the coming weeks.
- [Jan 2004] New Scientist,
"The World's No.1 Science & Technology Magazine", has a feature
article on the U of A computer poker research group in their 2003
year-end double issue (cover date: Dec 20 - Jan 9, 2004, pictured
above, pages 64-68).
- [Nov 2003] Vexbot, an experimental adaptive program
we are currently developing, won the gold medal at the 2003
Computer Olympiad held in Graz, Austria. Sparbot,
our game-theory based program, took the silver medal.
- [Nov 2003] The Daily Planet,
a science and technology news program on the Discovery channel,
did a segment on the U of A computer poker research group.
A
video clip of the segment is available on their website.
Darse was also interviewed in a
previously aired segment on the game of RoShamBo (aka
Rock-Paper-Scissors).
- [July 2003]
Sparbot has been added to the
available online games. This is a refined version of the
pseudo-optimal player (PsOpti-2 architecture), as described in
our paper for IJCAI 2003.
- [June 2003] There has been some media attention recently, including
an
NSERC news release, a front page story in the
Edmonton Journal, mention in a feature article in the
New York Times, and a
CBC radio interview.
|
|
Publications
|
- Strategy Evaluation in Extensive Games with Importance Sampling
Michael Bowling, Michael Johanson, Neil Burch, Duane Szafron, in Proceedings of
ICML 2008 (Twenty-Fifth Annual International Conference on Machine Learning).
[pdf]
- Using State Estimation for Dynamic
Agent Modelling
Nolan Bard, M.Sc. thesis, March 2008.
[pdf] [ps]
- Regret Minimization in Games with Incomplete Information
Martin Zinkevich, Michael Johanson, Michael Bowling, Carmelo Piccione, in Proceedings of
NIPS 2007, (Twenty-First Annual Converence on
Neural Information Processing Systems).
[pdf]
- Computing Robust Counter-Strategies
Michael Johanson, Martin Zinkevich, Michael Bowling, in Proceedings of
NIPS 2007, (Twenty-First Annual Converence on
Neural Information Processing Systems).
[pdf]
- Robust Strategies and
Counter-Strategies: Building a Champion Level Computer Poker Player.
Michael Johanson, M.Sc. thesis, October 2007.
[pdf]
- Postgame Analysis of Poker Decisions.
Morgan Kan, M.Sc. thesis, January 2007.
[postscript]
[pdf]
- Algorithms and Assessment in Computer Poker.
Darse Billings, Ph.D. dissertation, 2006.
[postscript]
[pdf]
- Optimal Unbiased Estimators For Evaluating Agent Performance.
Martin Zinkevich, Michael Bowling, Nolan Bard, Morgan Kan, and
Darse Billings, Proceedings of
AAAI-06, (Twenty-First National Conference on Artificial
Intelligence).
[postscript]
[pdf]
- A Tool for the Direct Assessment of Poker Decisions.
Darse Billings and Morgan Kan.
[postscript]
[pdf]
- Opponent Modelling and Search in Poker.
Terence Schauenberg, M.Sc. thesis, 2006.
[postscript]
[pdf]
- Here is a brief overview of PsOpti
to help understand the benefits and limitations of a game theory solution
for 2-player Texas Hold'em. It is an excerpt of a reply to a science writer
for the German newsmagazine DER SPIEGEL.
-
Approximating Game-Theoretic Optimal Strategies for Full-scale Poker.
Darse Billings, Neil Burch, Aaron Davidson, Robert Holte,
Jonathan Schaeffer, Terence Schauenberg, and Duane Szafron,
Proceedings of IJCAI-03, (Eighteenth International Joint Conference on Artificial
Intelligence), 2003.
(Winner of the IJCAI / AAAI 2003 Distinguished Paper Award)
[postscript]
[pdf]
-
The Challenge of Poker.
Darse Billings, Aaron Davidson, Jonathan Schaeffer,
and Duane Szafron,
Artificial Intelligence Journal, vol 134(1-2), pp 201-240, 2002.
[postscript]
[pdf]
-
Opponent Modeling in Poker: Learning and Acting in a
Hostile and Uncertain Environment.
Aaron Davidson, M.Sc. thesis, 2002.
[postscript]
[pdf]
- Improved Opponent Modeling in Poker.
Aaron Davidson, Darse Billings, Jonathan Schaeffer, and Duane Szafron,
Proceedings of the 2000 International Conference on Artificial
Intelligence (ICAI'2000), pp 1467-1473, 2000.
[postscript]
[pdf]
- Using Probabilistic Knowledge and Simulation to play Poker.
Darse Billings, Lourdes Peña, Jonathan Schaeffer,
and Duane Szafron, Proceedings of
AAAI-99, (Sixteenth National Conference of the
American Association for Artificial Intelligence), 1999.
[postscript]
[pdf]
- Learning to Play Strong Poker.
Jonathan Schaeffer, Darse Billings, Lourdes Peña,
and Duane Szafron,
ICML-99,
Workshop at the the Sixteenth International Conference on
Machine Learning, 1999.
[postscript]
[pdf]
- Using Selective-Sampling Simulations in Poker.
Darse Billings, Denis Papp, Lourdes Peña, Jonathan Schaeffer,
and Duane Szafron,
AAAISS-99,
Proceedings of the AAAI Spring Symposium on Search Techniques for
Problem Solving under Uncertainty and Incomplete Information, 1999.
[postscript]
[pdf]
- Opponent Modeling in Poker.
Darse Billings, Denis Papp, Jonathan Schaeffer, and Duane Szafron,
Proceedings of
AAAI-98
(15th National AAAI Conference), 1998.
[postscript]
[pdf]
-
Poker as an Experimental Testbed for Artificial Intelligence Research.
Darse Billings, Denis Papp, Jonathan Schaeffer, and Duane Szafron,
Proceedings of AI'98, (Canadian Society for Computational Studies
in Intelligence), 1998.
[postscript]
[pdf]
-
Probabilities and Simulations in Poker.
[postscript]
Lourdes Peña, M.Sc. thesis, 1999.
-
Dealing with Imperfect Information in Poker.
[postscript]
Denis Papp, M.Sc. thesis, 1998.
-
Computer Poker.
Darse Billings, M.Sc. research essay, 1995.
[postscript]
[pdf]
|
|
|