GT-NO: The Perfect Strategy… For A Computer

For the last several years, the world’s best chess, checkers, backgammon, and go player has been a computer. And it will almost certainly stay that way forever. But what about poker?

Those other games have no information hidden and psychology plays little or no part in correct strategy.

To put it simply, there is no bluffing.

But of course, there is plenty of bluffing in poker. Given that, how could a computer be the world’s best at that game?

It seems impossible, but it’s not. Because there is a technique that was developed by someone with a brilliant mind that a computer can adopt, that will ensure that the computer cannot lose money in the long run to any poker player, even though it is doing nothing to try to figure out how that player plays.

And presently there are in fact computers that are almost perfect at using this technique in two forms of poker: no-limit hold’em and limit hold’em. No human can beat them in the long run playing heads up at either game.

The brilliant mind I was speaking of is now more often called “a beautiful mind” because that was the title of the Academy Award-winning movie about him, and the technique that John Nash developed is often called “game theory optimum.” You’ve undoubtedly heard it called GTO, a term thrown around every other minute on high-stakes poker streams.

GTO is amazing because it does not lose in the long run to any counter strategy, as long as the game is fair and there is only one opponent. The worst it does is sometimes tie. It doesn’t matter whether the other player is loose or tight, passive or aggressive, or bluffs too much or bluffs too little. The computer is going to eventually beat him (or anyone else who doesn’t play GTO strategy).

When there is more than one opponent GTO can theoretically have a disadvantage, but in real life it doesn’t. Also, GTO strategy has not been derived close to perfectly for any poker games except for hold’em and some forms of five card draw.

GTO has a few interesting attributes.

1. It does not try to figure out how an opponent is playing. Basically, it comes up with a strategy that assumes that the opponent is also playing GTO. (Although there are now computers that can be asked to assume he plays differently, but that is no longer technically GTO.)

2. It does not make plays specifically designed to give a false impression of how it will play in the future.

3. It will mix up some of its plays by playing the same hand in different ways. But those ways will come at a mathematically-derived frequency via a random number generator. It has nothing to do with the proclivities of its opponents.

4. It will still win even if its exact strategy is divulged to its opponent!

5. It will usually bluff and make other type bets (and calls) just often enough such that when you can play your hand in more than one way, you won’t have a clearcut decision. (They call this “making you indifferent.”)

6. It will usually (but not always) call at least often enough such that you would not profit by betting two blank cards.

Another attribute is that the strategy can be to some degree memorized, so you don’t have to do much thinking on the fly. You only need to realize that a situation you encounter is close to one you studied at home and then do what the computer told you to do in that situation.

Thus, it is no wonder that a lot of players, including many of the world’s best, are gravitating towards the strategy of remaining as close to GTO as possible except in the very obvious cases where you should deviate. (For example, don’t use GTO bluffing frequencies against someone who never folds.)

Play as close to the strategy as you can, one that doesn’t require lots of spur of the moment thought, and eventually you are guaranteed to win. A game plan almost as brilliant as Nash himself, right?

Nope! It’s actually a terrible idea… at least for most of the games that the average reader will be playing in. As it turns out, a GTO strategy against a table full of amateur or low-stakes players might be profitable, but it is also leaving a lot of money on the table.

By learning when to deviate from GTO, and how to spot exploits in your opponents, you can maximize your win rate. I will give specific examples next time. ♠

David Sklansky is the author of The Theory of Poker, as well as nearly two dozen other guides on gambling, poker, and other games. The three-time WSOP bracelet winner’s latest book, Small Stakes No-Limit Hold’em: Help Them Give You Their Money, is now available on Amazon. You can contact Sklansky at [email protected].