Nearly a decade ago, researchers at the Computer Poker Research Group (CPRG) of the University of Alberta, Canada developed an artificially intelligent automaton to play Texas Hold’em. Dubbed “Vex Bot” because of its ability to frustrate human players, it was capable of playing poker at the master level. The only problem then was that its gambling genius could only be applied to heads-up games.
More important than the machine itself, however, was the thinking behind it. In games like chess or blackjack, developing optimal strategies is all about number-crunching. Poker, on the other hand, involves bluffing and laying traps for opponents. A huge amount of information is unavailable, so a layer of artificial intelligence (AI) is required in addition to the capacity for calculating probabilities.
Until 2008, the poker bots were good, but not very good. As New York Times reporter Gabriel Dance put it: “Humans were better at the nuances of the game—at bluffing, for instance—and could routinely beat the machines.” By 2011, that had changed, said Dance, to where “poker bots are now good enough to win tens of thousands of dollars on major game sites.”
A more recent poker showdown between professional players and an artificial intelligence program has ended with a slim victory for the humans. Although human players bested the AI software, the contest was close enough to be called a statistical tie. At the end of the event, combined winnings from the four humans were $732,713 over that of their digital opponent.
Four pros — Bjorn Li, Doug Polk, Dong Kim and Jason Les — settled down at Rivers Casino in Pittsburgh to play a total of 80,000 hands of Heads-Up, No-Limit Texas Hold ’em with Claudico, a poker-playing bot made by Carnegie Mellon University computer science researchers. This marked the first time that a computer program has played human professionals in a heads-up, no-limit poker game.
But the university has plans to update the supercomputer’s algorithms, and now a computer poker program called Baby Tartanian 8 continued Carnegie Mellon’s hot streak at the Annual Computer Poker Competition, taking first place in the total bankroll category and third place in the bankroll instant run-off category in the Heads-Up, No-Limit Texas Hold’em game. In the total bankroll category, the winner is determined by total winnings. In bankroll instant run-off, the pokerbot with the lowest total bankroll is eliminated in each round.
Baby Tartanian 8 was designed by Noam Brown, a Ph.D. student in the School of Computer Science, alongside his adviser, Tuomas Sandholm, a professor in the Computer Science Department. The project is representative of an incomplete information problem. The aim of this problem is to be able to find and play the strategy with the best results, based off limited information. Tartanian 8 was developed as a successor to the previous bot Tartanian 7 using the Comet supercomputer at the San Diego Supercomputer Center. Tartanian 8 was built entirely from scratch, boasting better, faster algorithms, and hardware that supports increased computation power. The algorithms were more advanced than those used to create Tartanian 7 and Claudico.
Although Tartanian 8 was built from the ground up it was developed as a successor to Claudico. Researchers have been able to analyse the data gathered during the Brains vs. Artificial Intelligence match up, and come up with a list of strengths and weaknesses that have been refined to create the new poker bot. The strengths were that Tartanian 7 often did things that humans wouldn’t. For instance, it would bet a very large sum of money on a very small pot or a very tiny amount to a very large pot. As opponents, they said that such strategies threw them off. According to them, that was one instance when the bot was unpredictable, as human players generally don’t make such decision, but the bot was unfazed, and was able to balance that move with other moves. The researchers also identified some of the weaknesses of the bot. One of the major weaknesses of the bot was that the algorithm was designed to put card hands into similar sets of hands on which it could use a similar strategy. This is called abstraction. When grouping those hands together, the bot would sometimes misjudge the situation, which would cause it to have it a suboptimal strategy and act absurdly.
The ultimate goal in this line of research is to beat the top humans, Brown said and Pokerbots will beat the top humans in the next 2-3 years, Tartarian 8 creator also says. Whether it happens in two – three years or more, almost all AI observers seems to agree that it will one day happen.