Ari Engel on Whether Limit Hold'em Is "Solvable"
Efforts to "solve" different poker variants have been ongoing for years, with heads-up, fixed-limit hold'em being the game researchers have focused on the most as they try to create computer-driven poker-playing programs that can provide human players a significant challenge.
We've been reporting on the quest to "solve" LHE for more than a decade here at PokerNews, dating back to Phil Laak taking on an early, computer-driven challenger Poker Probot in 2005. After that came the first "Man-Machine Poker Championship" that pitted a program built by the Universtiy of Alberta's Computer Poker Research Group against Laak and Ali Eslami in 2007.
The humans won those matches, but then the Alberta group's Polaris 2.0 beat a team of humans in 2008, and from there the battle continued.
More recently came a claim from another group at Alberta they have "solved" the variant, as described in the January 2015 article Five Scientists Claim They Have Solved Heads-Up Limit Texas Hold'em. That said, despite continued advances, including those demonstrated at the Annual Computer Poker Competition, building a program to beat no-limit hold'em remains a significant challenge, as described in this March 2016 article Bot Creator Says 'No Chance That No-Limit Texas Hold'em Is Going to Be Solved.
Fixed-limit hold'em remains a popular game in many poker rooms and online, with three LHE events on the 2016 World Series of Poker schedule. Picking up on a recent discussion on the Primetime Poker Report with Donnie Peters, our Sarah Herring spoke with poker pro Ari Engel about the current state of limit hold'em and whether or not advances in AI and what the researchers have come up with have affected the game all that much when played between humans.
Take a look:
For more on how far the development of artificial intelligence has come in the effort to create better poker-playing programs, check out Nikolai Yakovenko's "Artificial Intelligence and Hold'em" series:
- Artificial Intelligence and Hold'em, Part 1: Counter-Factual Regret Minimization
- Artificial Intelligence and Hold'em, Part 2: Clustering & "Solving" Heads-Up Hold'em
- Artificial Intelligence and Hold'em, Part 3: No-Limit Hold'em, The Next Frontier
Also see Yakovenko's report from the most recent competition (in which he competed), Poker and AI: Reporting from the 2016 Annual Computer Poker Competition.