The sharpest edge
at the table.

A counterfactual regret minimization solver plays each spot against itself over and over. After every pass it looks back at each decision and asks how much it regrets not having chosen something else, then shifts weight toward the actions it regrets skipping. Repeat long enough and the strategy stops moving: neither side can improve by changing anything. That fixed point is the equilibrium: the baseline GTO play.

No-Limit Hold’em cash, solved at the depth you actually sit down with. One set of assumptions across the whole app, so preflop and flop never disagree about the pot.

A solved strategy is only valid for the bet sizes it was solved with, so the app’s buttons use exactly these. The size you tap is the size the solver priced.

Twelve real preflop lines are carried to the flop (six per table size, four single-raised pots and two 3-bet pots each) and solved through to the river. Turn and river carry a single bet size on purpose: it keeps the tree small enough to converge properly, which is what makes the flop answer trustworthy.

All 22,100 possible flops collapse to 1,755 strategically distinct boards once you ignore which suit is which. From those, 95 were drawn by a stratified sample across pairedness, suitedness, high card and connectivity, each weighted by how often it genuinely comes off the deck, so the boards you’re dealt match the boards you’ll actually face.

Each flop solve ran to an exploitability target of 0.5% of the pot, meaning the most a perfect counter-strategy could win against it. What lands on your phone is the full solution: per-hand action frequencies and per-hand EVs, roughly 20 MB, checked against a stored fingerprint on load. Grading happens on-device against those exact numbers.