Viewers Guide: Poker Pros Take On Poker Computer In ‘Brains vs. AI’
Doug Polk, aka “WCGRider,” and Dong Kim, aka “Donger Kim,” are part of a four-person team that will attempt to defeat a computer program at poker over a two-week event going on now.
The competition is called “Brains vs. Artificial Intelligence,” and it was devised by Dr. Tuomas W. Sandholm, Professor of Computer Science at Carnegie Mellon University in Pittsburgh, Pennsylvania. Sandholm is the developer of an AI program called “Claudico” that plays heads-up no-limit Texas hold’em. At the recently held “Annual Computer Poker Competition,” Claudico defeated a number of other poker-playing computers.
Now, Sandholm is putting his creation to the test against human players in a two-week competition, that is being shown live on the internet.
It’s hardly the first time a computer has been programmed to play poker, nor the first time humans have tried to defeat a computer opponent at poker. But this is the most serious attempt to see how a computer stacks up against some of the best pros in the world, with a significant sample size.
Read on for more details about the contest.
Who is playing?
The lineup of pros is a solid one. The selections erred on the side of players who know game-theory optimal play, as opposed to more famous pros.
–The most famous of the lot is Polk. He is an online beast who plays as WCGRider. He also has a World Series of Poker bracelet and more than $3 million in live tournament earnings.
–Kim is also known for his online play, as “Donger Kim.” He specializes in heads-up play, and recently won a heads-up challenge that was followed by the online poker community.
–Bjorn Li is a pro with a degree in mathematics from the University of Chicago. He also specializes in heads-up play, and has half a million dollars in live cashes. His online handle is unknown.
–Jason Les holds a degree in computer science from UC-Irvine and is an online and live pro. He has half a million dollars in live cashes.His online handle is unknown.
When is the contest?
The contest began on April 24th, and it will continue through May 7th. Play will begin around 11 a.m. Eastern every day, and will conclude around 10 p.m., with some room for variance based on the schedules of the pros. There will be no play on May 3rd. The official results will be revealed on May 8th.
So how does the contest work?
All four pros will play about 1,500 hands per day against Claudico over the course of the competition. A total of 80,000 hands will be played.
The pros are at Rivers Casino in Pittsburgh. They are playing on laptops connected to a computer at Carnegie Mellon. From the official website of the competition:
Two pros will play on the casino main floor and two will play in an isolation room on the second floor. To reduce the role of luck, the pros in the isolation room will play the opposite hole cards against Claudico from the ones being played by the pros and Claudico on the main floor. The players will rotate periodically between the main floor and isolation room.
The winner of the competition will be whoever has the most chips after 80,000 hands: the humans or Claudico.
So I can watch this?
Yep, it’s being shown, in its entirety, on Twitch. All action is on a 30-minute delay.
We’ll include Polk’s feed below, and then link the others, since the sound autoplays. It’s not just watching pros clicking buttons, since they are allowed to interact with one of the other players.
If you want to watch them simultaneously on one feed, you can do that too. Click here for the multifeed. There is no sound there.
A scoreboard is also being kept live on Twitch, with periodic updates:
How good is the computer?
Claudico has been under development for nearly 11 years.
From the press release on the competition:
“Claudico” has been designed to play optimally – at all times. Every strategic decision made by the computer is based on a complex algorithm which has steadily been revised and improved by a team of math and computer experts at CMU. Now, Dr. Sandholm believes his perfect poker player constituted in form of a software program is ready to face the toughest human competition. And now comes the ultimate test – actual face-to-face competition in a live cardroom.
From the research team:
Claudico’s strategy was computed purely algorithmically by Prof. Sandholm and his PhD students Noam Brown and Sam Ganzfried using the Blacklight supercomputer at the Pittsburgh Supercomputing Center. Around 2-3 million core hours were used to compute the strategy.
Also, from the competition site:
Claudico performs real-time reasoning during a hand. It also improves its strategy during the match by computing continuously on the Pittsburgh Supercomputing Center’s Blacklight computer. The research team built Claudico using algorithms that analyzed poker rules to devise a winning strategy. Claudico is not based on the experience of expert human players, so its strategy for playing can differ markedly from seasoned pros.