Train your skillsLearn a new poker variantFind out how well you actually play
Current Overall Standings:Hands played: | 15563 | |
Human Winnings: | -2907 $ | (in imaginary Dollars) |
Human Winrate: | -9.3 bb/100 | (Big Blinds / 100 Hands) |
I created a bot / an algorithm / an artificial intelligence - I don't care how you prefer to call it - which plays Deuce To Seven Heads Up Poker. You can play against it here:
This project is still in a very early stage, so please don't be surprised if something is not working properly or if you encounter an error. I'd love to hear your feedback!
I'm interested in finding out how good / bad the algorithm actually plays and how it compares to human players. In case you are interested, the bot's strategy has been precomputed and is fixed. It never changes no matter how you play. But that does not mean that it always makes the same decision in the same situation. Its strategy in a specific situation might be to take one action 70% of the time, and another action 30% of the time. The strategy was created via a technique called counterfactual regret minimization, which is basically a fancy version of letting the computer play against itself and improving over time. There is plenty of academic literature about how it works. A good starting point can be found here.