Little GTO

There's an interesting thread on 2+2 about whether the top poker pros use game theory as part of their strategy. It appears that some do while others are blessed with an innate ability that acts as a substitute.

Today there are bots employing GTO (game theory optimal) strategies that could beat any human in heads-up limit hold'em (LHE). In the beginning of the thread, no less than Durrrr (Tom Dwan) issued a challenge whereby he was willing to take on any bot in LHE. There were a few takers but as of yet, Durrrr has not agreed.

No limit hold'em (NLHE), with its different betting structure, is a much more complex endeavor when it comes to programming a GTO bot. In fact, the largest game ever solved was checkers (2007), which took a team led by Jonathan Schaeffer nearly 16 years and involved 200 desktop computers. NLHE is more complicated than checkers.

Incidentally, Professor Schaeffer had this to say about solving chess :

"The number of chess positions possible (as calculated in several chess books, a couple of which are chess encyclopedias which probably quote from other sources) is greater than the number of atoms in the universe (assuming that it is 10-15 billion light-years in diameter). So roughly, the complete game tree would not fit on a computer the size of the universe. But, the entire game tree does not have to be on the computer at once. And many positions are not important. Some experts in artificial intelligence think that a very large computer (larger than any in existence, but certainly smaller than a planet) may be able to solve chess by calculating for a few thousand years, maybe for hundreds of years."