Equity is the most important single number in poker decision-making. Every +EV bet depends on it. Your equity is your share of the pot if the current street runs out with no more betting.
Two kinds of equity:
1. Equity against a specific hand. If you have A-K and your opponent has K-Q, you have about 73 percent equity (you win 73 percent of the time). This is what an equity calculator tells you when you plug in two specific hands.
2. Equity against a range. Your opponent does not show you their cards, so your real equity is against the range of hands they could hold given the action so far. Against a tight opener range of {TT+, AK, AQs}, AK has about 35 percent equity. Against a loose opener range of {22+, Ax, KQ, QJ, suited connectors}, AK has about 55 percent equity. Same hand, different equity, different decision.
Why range-based equity matters:
Beginner bot developers plug their hand into an equity calculator against a single opponent hand and get misleading numbers. The right framework is: your hand vs. opponent range, weighted by betting action.
Example: You have AKo on a T-7-2 rainbow flop. Your opponent check-raised you. What is your equity?
- Against a value-only range of {TT, 77, 22, T7s, AT}: roughly 12 percent. You are crushed.
- Against a balanced range that includes bluffs like {T7s, AT, TT, 77, A5s semi-bluff, 98s semi-bluff}: roughly 35 percent.
- Against a pure bluff range: roughly 50 percent (six outs to improve).
Same hand, same board, same action, but the equity range is 12 percent to 50 percent depending on what you think your opponent has. That is why opponent modeling matters. Without a read, your equity estimate is a guess.
Computing equity in your bot:
You do not need a full equity calculator. A simple bucketed approach works for most heuristic bots:
- Pre-flop: use a static hand-strength table (sklansky rankings, or a custom weighting).
- Postflop: count your outs (cards that improve your hand to a winner) and use the rule of 4 and 2: multiply outs by 4 on the flop for equity to the river, by 2 on the turn for equity to the river.
More accurate methods (Monte Carlo simulation, precomputed range vs. range equity tables) improve accuracy but add complexity. For a bot running on Open Poker, a simple outs-based estimator is enough to beat most of the field.