PulseGTO
Same small sizing. Different checking ranges.
Full frequencies and calculation settings below. These are three examples from one custom configuration.
A small bet can be a solver's most-used sizing without being an action it takes with every hand. Those are two different observations: how much the solver bets when it bets, and how often it bets at all.
We examined three rainbow flops in the same BTN-versus-BB single-raised-pot setup. A quarter-pot bet was the most-used individual bet size on all three. Yet BTN checked between 31% and 48% of its range after BB checked.
That does not establish that betting everything small would be an expensive simplification. We did not measure that cost. It establishes something narrower: these computed strategies retain substantial checking ranges.
The spot
The game starts six-handed with 100 bb stacks. Action folds to BTN, who opens to 2.5 bb. SB folds and BB calls. The flop pot is 5.5 bb, with 97.5 bb remaining for each player.
Both preflop ranges come from one custom 6-max tree with 5% rake, a 4.5 bb cap and no flop no drop. They are the ranges reaching this particular line, not generic charts for every game.
Each flop was re-solved from BB's first decision. BB could check or lead; the BTN results below are for the decision after BB checks. BTN could check or use nominal 25%, 50% and 75% bets, with all-ins enabled in the configuration. The recorded BTN action menu at this node contains check and those three bet sizes.
Three dry flops, three mixtures
| Flop | Check | Bet 25% | Bet 50% | Bet 75% |
|---|---|---|---|---|
| A♠7♦2♣ | 44.5% | 46.4% | 3.6% | 5.5% |
| K♠8♦3♣ | 31.4% | 51.9% | 9.7% | 7.1% |
| Q♠7♦4♣ | 48.3% | 24.5% | 11.2% | 16.0% |
Range-weighted frequencies, rounded to one decimal; totals may differ slightly from 100%.
On A72, the quarter-pot bet accounts for about 84% of BTN's bets, but only 46% of its entire range. Almost 45% checks back.
K83 has the highest overall betting frequency of these three examples: approximately 69%. The small bet is prominent, but BTN still checks almost one hand in three.
Q74 illustrates why “most-used sizing” needs care. The quarter-pot bet is the largest individual bet category at 24.5%, but the two larger bets together account for 27.2%. Almost half the range checks.
Look inside the checking range
The range-level result becomes more useful when we inspect individual combinations. These are examples from the saved solve, not instructions to reproduce decimal frequencies at the table.
On A♠7♦2♣, A♣3♣ checks about 45% and bets quarter pot about 53%. Having top pair does not automatically put a hand into an always-bet category. K♦8♦, an unpaired hand with a backdoor diamond flush possibility, checks about 96%.
On K♠8♦3♣, A♥7♥ checks about 99%. K♥7♥, a weak top pair, checks about 51% and bets quarter pot about 45%.
On Q♠7♦4♣, A♥9♥ checks essentially always. Q♥3♥ checks about 94%, while 4♦4♥ never checks in the recorded strategy and distributes its bets across all three sizes.
These observations show that checking is not restricted to one simple category such as “complete misses.” Explaining every combination's exact mix would require more than its hand strength label: suits, blockers, opponent ranges and future branches all enter the model. The examples demonstrate the mixture; they do not isolate its causes.
What this changes in a study session
When you see a solver use a small c-bet, ask two questions before turning it into a shortcut:
- Is small betting frequent across the entire range, or simply frequent among the hands that bet?
- Which combinations remain in the checking range?
The second question is often more revealing than memorizing the main sizing. Here, substantial checking remains on all three boards even though the small bet is the most-used individual bet size.
What we have not shown
These solves do not measure the EV loss from forcing every hand to bet 25%. That requires a separate restricted-strategy evaluation with the opponent allowed to respond appropriately. Aggregate EV tiles and the frequency of checking do not supply that answer.
The results also do not establish a rule for all dry flops, different opening sizes, different rake structures or a particular opponent pool.
Reproduction notes
- Three separate heads-up postflop RangeCFR re-solves in PulseGTO, from a custom 6-max preflop tree.
- Ranges below the GUI's default relative cutoff of 1% of the largest combo reach were omitted.
- BB's initial lead option: 50% pot. BTN's nominal flop bets: 25%, 50%, 75% pot.
- Actual initial BTN bet amounts, rounded to 0.01 bb: 1.38 bb, 2.75 bb and 4.12 bb — approximately 25.1%, 50.0% and 74.9% of 5.5 bb.
- Turn and river bets: 75%; OOP donk options: 50%. Raises: 3× on flop/turn, 2.5× on river, maximum two raises per street; all-ins enabled. These are simplified downstream trees.
- Reported final accuracy: A72 0.271%, K83 approximately 0.300%, Q74 0.286% of the starting pot. This is the engine's reported average best-response gap across both players, not a guarantee of each individual action frequency's precision.
- Results are finite-accuracy solutions for these ranges and trees, not exact universal GTO frequencies. Changes in tree options can change the mixtures.
PulseGTO was used for the calculations and recorded outputs. The purpose of the exercise is to show how to inspect a strategy rather than infer an entire game plan from its most visible sizing. Explore the postflop solver.