What is GTO poker?
GTO (game theory optimal) poker means playing a Nash equilibrium strategy: betting, calling and bluffing in proportions balanced so that no opponent can exploit you. It isn’t about reading anyone — it’s about being impossible to punish. Set a bluffing frequency on the felt and see why.
The river bluffing game
Pot 100. Half your river hands are the nuts, half are air. Your opponent holds a bluff-catcher: it beats every bluff and loses to every value bet.
Too few bluffs. A bluff-catcher profits by folding every time, so your value bets never get called: 62.5 chips per hand, 12.5 less than balanced.
Chips per hand
What GTO means, in plain English
GTO is short for game theory optimal. The idea comes from John Nash, who showed in 1950 that games like poker have an equilibrium: a set of strategies, one per player, where nobody can win more by changing only their own. A GTO strategy is your part of that equilibrium.
In poker, a strategy is a plan for every hand you could hold in every spot — including how often to do each thing. GTO play is usually mixed: the same hand might bet 60% of the time and check 40%. Mixing is what makes a strategy hard to read, because your actions never reveal your cards reliably.
The reward for all that balance is a guarantee. In heads-up poker, where the two players take turns in each seat, an equilibrium strategy can’t be beaten in the long run: whatever the opponent does, the best they can manage is to break even. With three or more players the math no longer promises that — other players’ choices can cost you even when you play an equilibrium — but strategies built on the same ideas still play very strongly.
Why a balanced bluff can’t be exploited
The game on the felt is the textbook case. You bet the river holding either the nuts or nothing. Your opponent holds a bluff-catcher, a hand that beats only your bluffs. The pot is 100 and you bet 100.
Your opponent risks 100 to win 200, so calling breaks even when a third of your bets are bluffs. Bluff more than that and calling every time wins them chips; bluff less and folding every time does. At exactly one bluff for every two value bets, calling and folding earn them the same — so no reply helps them, and you collect 75 chips per hand whatever they do.
Move the slider off balance and watch the tent: an opponent who knows your frequency picks the reply that hurts you, and every step away from the peak costs chips. The mirror image holds for the caller. If they call half the time against a pot-sized bet, your bluffs break even and you can’t exploit them either. Both sides balanced, nobody wants to change: that’s the equilibrium.
The two GTO formulas: bluffing frequency and minimum defense
The river game generalizes to any bet size. Two numbers fall out of it:
Bluff share of a polarized betting range = bet ÷ (pot + 2 × bet)
Minimum defense frequency (MDF) = pot ÷ (pot + bet)
A pot-sized bet: bluffs are 100 ÷ 300 = 33.3% of bets, and the caller must continue 100 ÷ 200 = 50% of the time.
Bet size vs bluffs and defense
| Bet size | Bluffs in range | Value per bluff | Minimum defense | Folds a bluff needs |
|---|---|---|---|---|
| Quarter pot | 16.7% | 5.0 : 1 | 80.0% | 20.0% |
| Third pot | 20.0% | 4.0 : 1 | 75.0% | 25.0% |
| Half pot | 25.0% | 3.0 : 1 | 66.7% | 33.3% |
| Two-thirds pot | 28.6% | 2.5 : 1 | 60.0% | 40.0% |
| Three-quarters pot | 30.0% | 2.3 : 1 | 57.1% | 42.9% |
| Pot | 33.3% | 2.0 : 1 | 50.0% | 50.0% |
| 1.5× pot | 37.5% | 1.7 : 1 | 40.0% | 60.0% |
| 2× pot (overbet) | 40.0% | 1.5 : 1 | 33.3% | 66.7% |
Bigger bets allow more bluffs and let the caller fold more. Two cautions keep the formulas honest. They describe the last bet of a hand: on the flop and turn, a bluff usually has outs, so equilibrium strategies bluff more there. And MDF applies to your whole range — if your line really has no strong hands, defending by the formula just donates chips.
Unexploitable isn’t the same as maximally profitable
Look at the dashed lines on the felt. Against someone who always calls, never bluffing and betting only the nuts earns 100 chips per hand — a third more than the balanced 75. Against someone who always folds, bluffing with every weak hand earns 100 as well. The balanced strategy leaves that money on the table on purpose: it earns 75 no matter who sits across from you.
That is the trade at the heart of poker strategy. In the full game, an equilibrium strategy still profits from many opponent mistakes, but it never bends itself to punish a particular leak. Exploiting a leak wins more — and opens a leak of your own, because an opponent who notices can turn it against you.
GTO vs exploitative play
| GTO (equilibrium) play | Exploitative play | |
|---|---|---|
| Goal | Be impossible to exploit | Win the most against this opponent |
| Needs | Sound frequencies and balanced ranges | Reliable reads on how someone plays |
| Risk | Leaves profit behind against weak players | Can be counter-exploited if the read is wrong |
| Best when | Opponents are strong or unknown | Opponents have clear, repeated habits |
Strong players mix both. They start from a game-theory baseline, then shift toward exploitation once an opponent shows a pattern: fewer bluffs into someone who never folds, more pressure on someone who folds too often. At a home game with friends, most of the money is in those adjustments — see the types of poker players and how to play against each.
How GTO strategies are computed
Nobody works out a GTO strategy by hand for real poker. Programs do it with counterfactual regret minimization (CFR): the computer plays the game against itself over and over, and after each pass it asks, for every decision, how much better each alternative would have done. It shifts toward the actions it “regrets” not taking. In two-player games, the average of all those strategies converges on an equilibrium.
No-Limit Hold’em is far too big to compute whole, so every practical strategy is an approximation. Solver programs limit the bet sizes they consider, group similar hands together, or compute one spot at a time from assumed ranges and stack depths. Their outputs are excellent guides — and only as good as those assumptions. The story of how these methods beat professional players is on our poker AI page.
Common misconceptions about GTO
- “GTO gives one right move.” Equilibrium strategies often mix: the same hand bets some of the time and checks the rest. Several actions can be equally good.
- “Playing GTO means you won’t lose.” You lose plenty of hands — and whole sessions — to cards. The guarantee is about the long-run average in heads-up play, not any single night.
- “GTO means bluffing a lot.” Balanced ranges are mostly value. Even a double-pot overbet is only 40% bluffs; a half-pot bet is 25%.
- “A chart is the answer.” Every chart and solver output rests on assumptions about stacks, bet sizes and ranges. Change those and the strategy changes.
- “It’s only for pros.” The core habits — pricing calls with pot odds, bluffing with a sensible ratio, not folding too often to small bets — pay off at any table.
How FizzPoker uses game theory — only after the game
FizzPoker never whispers advice during a hand. Once a game ends, every decision a signed-in player made can be compared with game-theory models and labeled acceptable, inaccuracy, mistake or blunder by its estimated cost as a share of the pot (up to 10%, 10–30%, 30–80%, above 80%). That’s the Fizz Score: an advisory second opinion from simplified models, not the final word. Multiway spots are estimates, and unusual spots the models can’t judge are left ungraded and say so.
If you’d rather test yourself against game theory directly, Level 5 bots play from precomputed game-theory strategy tables where they exist — see poker against bots. And to size up a single hand, try the poker hand analyzer.
Questions people ask
What does GTO mean in poker?
GTO stands for game theory optimal. It describes a strategy that is part of a Nash equilibrium: a balanced way of betting, calling and bluffing that no opponent can exploit, because no change in their strategy makes them more money against it.
Is GTO the most profitable way to play poker?
No. GTO is the safest strategy, not the most profitable one. Against an opponent with a clear leak — someone who calls too much or folds too much — an exploitative adjustment wins more, at the cost of becoming exploitable yourself if you misread them.
How often should you bluff on the river in GTO?
With a polarized range (strong hands and bluffs), the balanced share of bluffs is bet ÷ (pot + 2 × bet). That is 33.3% of your bets for a pot-sized bet, 25% for a half-pot bet and 40% for a double-pot overbet.
What is minimum defense frequency (MDF)?
MDF is how often you must continue against a bet so that a pure bluff can’t profit: pot ÷ (pot + bet). Facing a pot-sized bet it is 50%, facing a half-pot bet 66.7%. It is a guideline for your whole range, not a rule for any single hand.
Can a human play perfect GTO poker?
No. Full No-Limit Hold’em is far too large for anyone to memorize an equilibrium, and even computer strategies are approximations built with simplified bet sizes. Humans use GTO as a baseline: sound frequencies, balanced ranges and pot-odds discipline.
Does FizzPoker give GTO advice during a hand?
No. FizzPoker never advises anyone while a hand is being played. After the game, signed-in players can see each graded decision compared with game-theory models and labeled acceptable, inaccuracy, mistake or blunder — an advisory second opinion, not a verdict.