GTO stands for Game Theory Optimal. It gets described as “the strongest way to play”, but the actual definition is narrower: a strategy no opponent can profit more from, however they adjust.
Those are two very different claims. Being unexploitable is not the same as winning the most. Confusing them is where most people take a wrong turn learning GTO.
- ✓GTO aims at not being exploited, not at extracting the most from each opponent.
- ✓Balance means the ratio of value hands to bluffs in one action leaves your opponent indifferent between calling and folding.
- ✓Betting half pot, bluffs make up about a third of the betting range. Facing half pot, you must defend about 67%.
- ✓Low-stakes opponents leak badly, so exploitative play earns more than balance. GTO’s role there is as a baseline.
- ✓A mixed strategy is not switching on instinct. It needs a randomisation method you can actually execute.
What GTO is actually solving
Start with a concrete spot. You bet the river. If your betting range holds only value hands and no bluffs at all, your opponent gets a free answer: fold every hand that cannot beat your value range, and never lose in the long run.
Reverse it. If your betting range is mostly bluffs, the answer is just as easy: call with anything that beats a bluff, and profit over time.
Both extremes have a correct response, which means both can be targeted. What GTO does is remove that correct response — it makes calling and folding worth exactly the same to your opponent. Once either choice is equally good, there is no direction left for them to exploit.
So GTO is a defensive floor, not an attacking weapon. It guarantees you cannot be broken, not that you earn the most. Understanding this stops you spending your study time in the wrong place.
Two numbers: bluff ratio and minimum defence frequency
Balance is not a feeling. It is calculated, and getting started only takes two formulas.
When you bet, how much of the range should be bluffs
To call, your opponent pays B (your bet) to win the pot P plus your bet B. The equity they need is:
B ÷ (P + B)
They only win when you are bluffing, so the proportion of bluffs in your betting range should equal exactly that number:
| Bet size | Bluff share | Value : bluff |
|---|---|---|
| One-third pot | 25% | 3:1 |
| Half pot | 33% | 2:1 |
| Pot | 50% | 1:1 |
In practice: the river pot is 100 and you bet 50. If you would bet six value combinations in that spot, you want roughly three bluff combinations alongside them. Bluffing too rarely is the most common imbalance at low stakes — the opponent never loses by folding, and your value hands never get called.
When you face a bet, how much you must defend
Now the other side. An opponent bets B into pot P. If you fold too often, they profit bluffing with any two cards. To make a pure bluff break even, you have to defend at least:
P ÷ (P + B)
| Their bet | Minimum defence | You may fold |
|---|---|---|
| One-third pot | 75% | 25% |
| Half pot | 67% | 33% |
| Pot | 50% | 50% |
This is the minimum defence frequency. Its value is not in defending the number to the decimal — it is in showing you that you fold too much, which is the biggest leak most beginners have.
At low stakes, learn exploitation before GTO
This is the piece of old advice that most needs updating.
GTO earns its value when opponents are also strong. When nobody has an obvious leak, the only edge left is not showing one yourself. At low stakes the situation is the opposite: the leaks are so large that playing a balanced strategy actively throws money away.
Two examples. An opponent who almost never folds to a bet: balance tells you to keep some bluff ratio, but the exploitative answer is stop bluffing entirely, bet only value, and size up. Another opponent folds far too often on the flop: balance calls for a balanced continuation betting range, and the answer against them is to bet any two cards. In both cases the targeted adjustment earns considerably more.
So the real relationship is this: balance is the baseline, exploitation is how you actually make money. You need to know what balanced looks like to see how far an opponent deviates, in which direction, and what to change. Skipping the baseline and adjusting at random is not exploitation — it is guessing.
A learning order: three things, in this sequence
Do not start by opening a solver and running complex spots. That burns a lot of time and produces very little you can use. This order pays back far better.
Step one: preflop ranges
Preflop is the one area you can memorise outright and use in every single hand. Fix your opening range for each position, plus your 3-betting and calling ranges facing a raise. For how position shapes those ranges, see preflop basics. Finishing this step removes over half your leaks, and it needs no calculation at the table.
Step two: continuation bet frequency and sizing
The flop carries the largest volume of decisions. What matters is not how to play each hand, but which boards call for frequent small bets and which call for infrequent large ones. Dry boards that favour the raiser — something like A♦8♣3♥ — suit a high frequency of small continuation bets. Wet boards that favour the defender need restraint.
Step three: choosing river bluffs
Back to the table above. The river is where imbalance costs the most, because the money is biggest and there is no later street to correct it.
Beyond the ratio there is the question of which hands to bluff with. Prefer combinations that have no showdown value and block your opponent’s strong hands. On a river board of K♠9♥4♦2♣7♠, Q♠J♠ makes a better bluff than A♦6♦: the first can never win at showdown and removes some of their KQ and KJ, while the ace-high occasionally wins the pot unimproved and wastes that value as a bluff.
How to actually execute a mixed strategy
Solvers routinely output something like “bet 70%, check 30%”. But at the table you only act once — so how do you execute 70%?
The answer is randomisation, not instinct. Two practical methods:
- Use the suits of your hand as the random source. With AQ suited, for example, decide in advance that the spade combination bets and the heart combination checks, giving you roughly an even split.
- Use the second hand of a clock. Glance at the seconds before you act and take action A if it falls in a given range.
More important though: you do not need mixing to start. Against low-stakes opponents, the cost of picking one pure strategy is very small — far smaller than the attention you would spend executing a mix. Mixed strategies become genuinely useful only once opponents start adjusting to your patterns.
Using a solver without wasting time
Solvers calculate the balanced solution for a given spot, but the answer assumes both players are playing perfectly. Copying it straight into a low-stakes game does not make you more money.
Three more effective ways to use one:
- Ask one question at a time. Something like “is my continuation bet frequency too high on this board type?” rather than dropping a whole hand in and staring at numbers.
- Learn the structure, not the digits. What is worth remembering is which board types call for more betting and which call for restraint — not that a particular node was 63.5%.
- Measure your own frequencies first. Pull your real tendencies from your hand history, then compare against the solution. That tells you what to fix.
