What is GTO poker?

GTO stands for Game Theory Optimal. A GTO strategy is a way of playing poker that cannot be beaten in the long run, no matter how your opponent adjusts. It is the poker application of a concept from game theory called a Nash equilibrium — a set of strategies where no player can improve their expected result by changing their own play alone.

In plain terms: if you play a GTO strategy, you have no weakness for an opponent to attack. They can study you forever and still not find a counter-strategy that wins money against you. The trade-off is that a GTO strategy does not actively try to punish their mistakes either — it simply refuses to make any of its own. (For when to deviate, see GTO vs exploitative play.)

A simple definition

Note

A GTO strategy is a Nash-equilibrium strategy for poker. Against it, every counter-strategy breaks even or loses — so it is unexploitable.

Poker is a game of imperfect information: you cannot see your opponent's cards. That hidden information is what makes poker hard to solve and is exactly what game theory was built to handle. A GTO solution tells you, for every possible situation, the mix of actions that keeps you balanced and unexploitable.

Why GTO play is "mixed"

A common misconception is that GTO poker gives you one correct move in each spot. It usually does not. Equilibrium strategies are typically mixed: the same hand in the same spot might bet 70% of the time and check 30% of the time.

This mixing is not indecision — it is the point. If you always took the same action with a given hand, an observant opponent could read you. By mixing in the right proportions, you stay unpredictable and balanced. A solver's job is to compute these exact frequencies for every hand in your range.

Pure action — always do one thing (e.g. always bet). Easy to read, often exploitable.
Mixed action — split between options at solver-computed frequencies. Balanced and unexploitable.

What a solver actually produces

A poker GTO solver takes a defined situation — stack depths, board, both players' ranges, and the available bet sizes — and computes the equilibrium strategy for every hand. The output is a full game tree of decisions with action frequencies and expected values at each node. You then study that solution to learn what balanced play looks like and to find where your own game leaks.

ART/GTO is a desktop poker GTO solver for No-Limit Hold'em. It computes these solutions locally on your CPU and — unlike most commercial solvers — writes the entire solved tree to an open, self-contained file you own and can re-open forever. See Build your own GTO solutions for why that matters.

How GTO became practical

GTO is not new as an idea — the underlying mathematics dates to John Nash in 1950. What changed is that algorithms and hardware finally caught up. The breakthrough was Counterfactual Regret Minimization (CFR) in 2007, which made it possible to approximate equilibria in games as large as poker. By 2015 researchers had essentially solved heads-up limit Hold'em, and by 2017–2019 solvers had reached superhuman strength in no-limit.

ART/GTO runs Discounted CFR (DCFR), a modern, faster member of that same algorithm family. The full lineage — who proved what, and when — is laid out in The research GTO is founded on.

Key takeaways

GTO = Game Theory Optimal = a Nash equilibrium for poker.
A GTO strategy is unexploitable: it has no weakness to attack.
Equilibrium play is usually mixed — actions at specific frequencies, not single "correct" moves.
A solver computes these frequencies for a defined spot; you study the result to fix your leaks.
The math is from 1950; the algorithms (CFR → DCFR) made it practical for poker after 2007.