BLOG

GTO vs exploitative poker: when to leave the equilibrium, and by how much

GTO MAX AI team · 4 October 2026 · 8 min read

The argument between GTO and exploitative players has been running for a decade and both sides are right about half of it. The equilibrium is the strategy nobody can beat. It is also the strategy that beats nobody by more than their own mistakes allow. Every big blind above that comes from a deviation, and every deviation is a bet that you have read the opponent correctly. This post is about when that bet is good.

1. What the equilibrium guarantees, and what it does not

A game-theory-optimal strategy cannot be exploited: whatever your opponent does, they cannot win more than the structure of the game gives them. It is a floor under your results. It is not a ceiling on them, and it is not the most profitable strategy against any particular person, because it was computed against an opponent who defends correctly. Against a player who folds to 70% of river bets, the equilibrium still bluffs at the frequency that makes a correct defender indifferent, and leaves the rest of that player's over-folding on the table.

The equilibrium is what you play when you know nothing about the opponent. Everything you learn about them is a licence to deviate, and the licence is worth exactly as much as the evidence behind it.

2. The number that decides a bluff

A bet of size b into a pot of p breaks even as a pure bluff when the opponent folds b / (p + b) of the time. Three-quarters pot needs 43%; a pot-size bet needs 50%; a third of the pot needs 25%. The equilibrium bluffs just enough that a correct defender is indifferent. Against a player who folds more than the break-even, every extra bluff is pure profit; against one who folds less, every bluff is a loss and your value bets get paid more. That single comparison, this player's fold frequency against this bet's break-even, decides most river deviations.

3. The number that decides a call

Facing a bet of b into p, you need b / (p + 2b) equity to call: 27% against three-quarters pot, 33% against a pot-size bet. The equilibrium defends enough combinations that the bettor's bluffs break even. Against a player who bluffs more than the equilibrium frequency, bluff-catchers that are folds in theory become calls; against one who never bluffs rivers, calls that are correct in theory become folds. Your hand's equity has not changed. The opponent's range has.

4. How big a deviation

Two things set the size. The first is distance: a player who folds 70% when the bet needs 43% is far from indifference, and the correct adjustment is large (bluff every candidate hand). A player at 48% is near it, and the adjustment is small. The second is confidence: 70% over 14 observations is a guess, 70% over 140 is a fact. A good adjustment scales with both, and a sound rule is to deviate little until the sample is real and fully once it is.

The classic mistake is the reverse: a full adjustment after three hands, because the three hands were memorable. Memory keeps the dramatic sample. Statistics keep all of them.

5. Exploits have a cost

Every deviation opens you to a counter-exploit. If you stop bluffing against a station, a station who notices can start folding. In practice most players do not notice, and most of those who notice do not adjust, which is why exploitation is profitable at all. But the cost is real against strong, attentive opponents, and against them the right amount of deviation is small. The exploit is sized by who is across from you, not by how sure you are of the theory.

6. What this looks like in a real-time assistant

GTO MAX AI shows two lines for every decision. The GTO line is the equilibrium for the spot, with frequencies. The exploit line starts from it and moves in the direction the opponent's profile justifies, by an amount that depends on their distance from indifference and the size of the sample, and it says why: "Bet 75%: GTO mixes 60/40, this player has folded to 71% of river bets over 18 hands, above the 43% that makes the bluff break even." Below twenty hands on a player, no exploit line is offered. That is the arithmetic above, done for every player at the table, every hand. How the profile behind it is built: player profiling.

7. A short list of deviations the numbers justify most often

Against a player who over-folds to continuation bets: bet the flop with your whole range at the smallest size.
Against a station: stop bluffing rivers, value bet thinner, size up with value.
Against a player who never check-raises: bet more hands for thin value in position; your bets are never punished.
Against a player who over-bluffs rivers: call down with bluff-catchers you would fold in theory, and check strong hands to induce.
Against a player who folds too much to 3-bets: widen your 3-bet bluffs with the hands that block their continuing range.

Each of those is a frequency compared with a break-even. None of them needs feel. They need a count.

GTO MAX AI: the RTA that knows the player.

GTO line, exploit line and the reason, live, on up to four tables, in Chrome. GGPoker, Natural8, ClubGG, PokerNow, and PokerStars in beta. Try it for 30 minutes on Telegram.

Get a 30-minute trial

Coaching and study tool. Use it where real-time assistance is allowed: study, home games, private clubs, play money. Most online sites forbid real-time assistance. Advice can be wrong; every decision is yours. Terms of use.

More from the blog

Longer reads: the poker RTA guide and all guides.