Solver Setup: Inputs, Convergence and Model Limits
Specify a two-player postflop problem
The current browser solver starts with two ranges, a flop, turn or river board, a starting pot and effective remaining stack. It does not compute preflop equilibrium ranges, multiway pots or tournament ICM. Its saved-solution workflow is not a PioSolver file import.
Set inputs before reading frequencies
- Choose OOP and IP ranges that actually reach the node. Presets are illustrative, not verified position-specific solutions.
- Select exactly three, four or five distinct board cards.
- Set pot and remaining effective stack in consistent units, then the allowed bet and raise sizes for each street.
- Configure rake percentage and cap to match the model you intend. The controls do not reproduce every room-specific fee rule.
- Build the tree and solve. Check the achieved convergence measure rather than assuming the requested target was reached.
Separate three kinds of uncertainty
Exploitability describes convergence within the modeled game. Sensitivity asks how outputs change when ranges, sizes or rake change. Statistical confidence concerns estimates drawn from samples. A low exploitability value does not prove that guessed opponent ranges are accurate, and a sensitivity comparison is not a statistical confidence interval.
Use results for offline study and review, not live assistance. Keep the inputs with the output so another review can reproduce the modeled question.