Skip to main content
RePoker
Back to Blog

RePoker Blog

How to Read GTO Frequencies

A GTO chart does not only say raise or fold. When a square is split, the percentage is how often that hand takes each action in that exact spot.

This guide explains what a frequency measures, how a mixed strategy differs from a pure one, and how to weight those percentages by combinations. The worked numbers come from one published chart: a six-handed, 10-big-blind button push-or-fold matrix already on GTO preflop charts for 6-max. The article does not invent solver outputs, expected-value gaps, or a chart for a different stack or game.

If you still need the shorthand behind labels such as 22+ or ATs+, see how to read poker range notation. If you need the 6-4-12 arithmetic, see how to count poker hand combinations.

What a frequency actually measures

A frequency is the share of the time a strategy takes one action with one hand class in one exact situation. If a chart says "raise 25%, fold 75%," that hand is raised in a quarter of otherwise identical spots and folded in the other three quarters. The frequencies of the actions that are still available add up to 100%.

That percentage is not:

  • the share of all starting hands in the range (that is a combo-weighted range size)
  • your session VPIP or a HUD 3-bet statistic
  • an instruction to raise the next hand you see and fold later ones in a fixed pattern

A "25% raise" is the question many people type when they first see a split cell. The same reading applies to any other split: 82.7% shove and 17.3% fold, or 56.4% shove and 43.6% fold.

Pure strategies and mixed strategies

A pure action is taken every time. On the published 10bb button chart, shoves 100% of the time, and folds 100% of the time.

A mixed strategy takes more than one action with the same hand in the same spot. On that same chart, shoves 82.7% of the time and folds 17.3% of the time. Remembering only "suited connectors shove" erases that distinction, which is why the charts article already flags as a mixed hand.

Those three rows are not a general opening chart. They belong to this model only: six-handed, 10bb starting stacks, action folded to the button, a shove to 10bb or a fold, no ante, no per-hand rake, and no ICM. The frequencies come from RangeMyHand dataset v0.3.0, reproduced on RePoker under CC BY 4.0.

The percentage is for that hand, not for the grid

A cell that says 56.4% is not "shove 56.4% of all hands." It is the frequency for that hand class. The share of the full deck is smaller, and it depends on how many combinations the class has:

  • a pocket pair has 6 combinations
  • a suited hand has 4
  • an offsuit hand has 12

The published 10bb matrix states that the button shoves 33% of all 1,326 starting combinations. That figure is combo-weighted. A mixed offsuit hand moves the total more than a mixed suited hand with the same percentage, because it has three times as many combinations.

Hand in this 10bb button chartPublished frequencyCombinations in the classWeighted shove combinations
100% shove44.000
82.7% shove, 17.3% fold44 × 0.827 = 3.308
56.4% shove, 43.6% fold1212 × 0.564 = 6.768
46.4% shove, 53.6% fold41.856
0% shove (100% fold)120

’s percentage is lower than , but it contributes more than twice as many weighted shove combinations. The 3.308 figure for is already printed on the charts page; the other rows use the same method on the same matrix.

To count unweighted combinations for a list of hands, use the poker combo calculator. If a strategy mixes a class, multiply only after you have a published frequency. Do not invent one.

Why mixed hands show up at the edge of a range

A solver mixes a hand when its strategy is willing to take more than one action in that spot. If one action were clearly preferred every time, the output for that hand would be 100% or 0%. This article does not invent an expected-value gap for any hand.

You can see the pattern on the published matrix without adding new numbers. The suited-king row falls from at 85% to at 1.8%. The suited-connector row passes through at 84.5%, at 46.4%, and at 0.5%. The mixed cells sit where a pure shove region hands over to a pure fold region. Hands in the middle of a strong region, such as , stay pure.

That is also why a frequency is attached to one situation. Change the stack, the before-you action, the rake, or the payout model, and the same hand can become pure or mixed in a different way. A 100bb cash opening chart is a different question from this 10bb shove-or-fold chart.

How to use a mixed frequency in practice

You do not need a random-number generator to get value from a split cell. The useful information is often that the hand is marginal in this spot, not that you must hit 82.7% exactly.

A practical way to execute the published 10bb chart:

  1. Treat the extremes as pure. shoves 96.4% of the time in this matrix; for human execution that is an always-shove. shoves 1.8% of the time; that is an always-fold. Missing those tiny tails is a different kind of error from ignoring a hand that is mixed near the middle.
  2. Keep a rule for the middle mixes. In this matrix, that group includes hands such as (82.7%), (65.5%), (56.4%), and (46.4%). Those are the cells worth a consistent rule.
  3. Let the suits do the work when you want a split. A suited hand has four combinations, so one suit is 25%, two suits are 50%, and three suits are 75%. Two suits is a close practical stand-in for at 46.4%. For , two suits of the ace and a non-matching two give 2 × 3 = 6 of the 12 offsuit combinations, or 50%, which is close to the published 56.4%. This is an execution shortcut. It is not a claim about which suits the model preferred.
  4. Do not round every mixed hand upward. If every 46% or 56% shove becomes "always," the 33% shoving range quietly gets wider. Rounding is for execution; it is not a reason to promote every edge hand into the core of the range.

In a RePoker drill, first confirm what the exercise is testing. Recalling a published frequency and choosing an action are different tasks. The memorize guide already makes that distinction; this article only explains what the number means.

What a frequency does not decide

A frequency does not, by itself, tell you:

  • which chart to use (match stack, position, action, and game first)
  • whether a written range such as ATs+ is being mixed (notation names hands; it does not store frequencies)
  • the equity of one hand against a range
  • how an opponent will actually play

It is also not a player-wide VPIP number. The range chart editor already warns that the rates on a decision are for that decision, weighted by the hands that reach it.

If you do not have a published frequency for a hand, leave the class at 100% or exclude it. Do not fill the gap with a guess.

A short practice path

  1. Open the 10bb button push-or-fold matrix and find one pure shove, one pure fold, and one mixed hand—, , and are already labeled there.
  2. For the mixed hand, write the combo-weighted contribution (4 × 0.827 = 3.308 for ) so the percentage does not get confused with "how many squares are shaded."
  3. Inspect action frequencies for a 100bb cash decision in the range chart editor. Read the cell in the same way; do not import the 10bb percentages into that game.
  4. Practice the decision in the GTO preflop trainer. Use a later review session to check whether you still remember which hands were mixed.

FAQ

What does a 25% raise mean on a GTO chart? It means that, in that exact situation, the strategy raises that hand 25% of the time and takes the other listed action or actions the remaining 75% of the time. It does not mean that 25% of all starting hands are raises.

What is a mixed strategy in poker? It is a plan that takes more than one action with the same hand in the same spot, at stated frequencies. On the published 10bb button chart, is mixed: 82.7% shove, 17.3% fold. is not mixed: it shoves 100% of the time.

Do I need a random number generator? No. For practice, round very high and very low frequencies to always or never, and use a simple suit rule for the middle mixes. Consistency matters more than hitting a decimal.

Is the 10bb chart my cash-game opening chart? No. The worked frequencies in this article are for a six-handed, 10bb, button first-in shove-or-fold model with no ante and no ICM. A 100bb cash open or a big-blind defense uses a different chart.

Why is the range percentage different from the number of shaded squares? Squares are hand classes, and those classes are not equally likely. Weight each class by 6, 4, or 12 combinations and by its action frequency, then divide by 1,326. The charts article's FAQ covers the same arithmetic.

Should I memorize every mixed frequency? Not as a first step. Learn the situation and the clearly pure hands first, then add the mixed cells your study plan actually tests. See how to memorize poker ranges.

Can I copy frequencies from another site's chart? Only if the stack, action, rake, and payout assumptions match the game you are studying. A number without its model is not transferable. Do not invent a frequency to fill a gap.

Sources

The worked frequencies and the 33% overall shove rate are taken from the live RePoker page GTO Preflop Charts for 6-Max Cash and Short Stacks, which reproduces a six-max button 10bb first-in chart from RangeMyHand dataset v0.3.0 under CC BY 4.0. The live page states a source verification date of October 12, 2026. The identity 4 × 0.827 = 3.308 for is printed on that page. The other weighted rows in the table above use the same combination counts on the same published matrix.

Public explainers of the same reading task, used here as evidence that the question exists, not as a source of the numbers in this article:

Keep learning

Back to Blog

RePoker

Turn strategy concepts into table instinct.

Master preflop logic with 200+ minutes of video, lock in ranges with FSRS spaced repetition, and sharpen your reads in practice drills.

Featured on