Reading Short-Stack Push/Fold Charts
In brief: A shove chart and a calling chart answer different questions.
Core idea
A push/fold chart narrows the available choices. It is not a comparison of every small open, limp, and later action. A short stack does not automatically make every decision a push/fold decision.
The reference file `data/ante/pushfold-6.json` specifies 6 seats, blinds of 0.5/1bb, and a BB ante of 1bb. Stacks are measured before posting. Its depth list covers 5–20bb. That is the file's coverage, not a universal boundary for a playing rule. The numerical verification table checks the metadata and depth list directly.
The first player to shove can benefit when everyone folds. A player calling an existing shove faces a different cost and an already defined opposing range. Do not transfer a hand's color from the shove chart to the calling chart without checking the decision being represented.
Example hand
With A5s, first ask whether the earlier seats folded or somebody already shoved. For a first shove, inspect your seat's entry. For a response, identify both the shoving seat and your calling seat. This article does not assign A5s a numerical frequency: finding the exact condition is the exercise.
Repeat with a small pair. Feeling ahead of part of an opposing range is not enough to settle a call. Compare the action with folding under the stated conditions. Do not use a single observed result to evaluate the whole range.
Common mistakes
Do not substitute heads-up data for a multi-seat chart. Do not silently round an unavailable depth into a nearby answer. Check what EV means: the metadata describes the difference between the action and folding, in bb.
Model limits
After a shove, later seats call or fold, and the model ends the action at the first caller. Card removal is ignored. The calculation therefore does not cover a full game with additional callers or every possible continuation. A small error inside the reduced model does not extend its coverage.
Practice
Use the Study screen and quiz in the Practice tab. If the interface offers only heads-up questions, practice that distinction without recording a multi-seat assessment. Choose once per problem, read the explanation, and do not carry stacks between problems. Do not use these materials during real-money play.