What Is GTO Poker?

GTO stands for game theory optimal. In poker, it refers to equilibrium-based strategy within a specified game model. The available actions, number of players and payoffs determine what that solution means.

What does "unexploitable" actually mean?

In a two-player zero-sum game, an exact equilibrium strategy protects its game value against any opposing strategy. This is an expected result over the whole strategy, not a promise to win each hand or to break even from every position. Multiplayer poker does not generally have the same safety guarantee.

See CMU's explanation of multiplayer poker and our solver methodology and limitations.

GTO and the Nash equilibrium

GTO solutions in poker are closely tied to several game theory concepts. In game theory, the concept of a Nash equilibrium is a scenario where, assuming that all other players' strategies are fixed, no player could gain more by changing their own strategy. This results in an equilibrium state, where each player is playing the best possible strategy available to them, with no unilateral deviation improving their own payoff. In poker, it is this equilibrium state that is the "GTO" solution.

Changing the action tree changes the game being solved. A shove/fold solution does not tell you whether a limp, smaller raise or postflop line earns more. Treat it as a baseline for the stated decisions, rather than a solution to every option at the table.

What a GTO push/fold range looks like

When the only two options are shove and fold, a GTO strategy is a list of the 169 starting hands with a shove frequency attached to each one. Most of them are settled - 100% shove or 100% fold, and no interesting decision to make. The hands worth studying are the ones in between.

A mixed hand uses both actions some of the time. In an exact equilibrium, actions used with positive probability have equal value at that decision. Published frequencies come from approximate solves: sampling, convergence and rounding can also affect a displayed mixture.

Does GTO win the most money?

Not necessarily. A baseline solved against modeled opponents can leave value unclaimed when a real opponent makes a predictable mistake. A best-response solve targets that particular tendency; its result is only as useful as the assumptions behind it.

Against someone who folds far too often to a shove, the highest-EV response is not to stay balanced - it is to shove wider and collect the folds. That deliberately unbalanced range wins more than GTO does against that specific player, but can lose its advantage if they adjust. Which one you should be playing, and what counts as enough evidence to switch, is the subject of GTO vs. exploit.

Do you need to play GTO?

Studying a baseline helps you understand how position, stack depth and opponent ranges change a decision. It also makes an adjustment explicit: identify which assumption differs at your table, then compare the available actions.

Short stacks make shove/fold study useful, but they do not erase every other action. Smaller raises and limps can still matter. Use these charts for their stated restricted game and compare a fuller model when you need those alternatives.