Reading opponent stats, and when to trust them

In this lesson

• Read each stat in the opponents panel and the situations it counts.
• Explain why a frequency over a few hands is mostly noise.
• Tell a stat that settles quickly from one that needs hundreds of chances.
• Turn a stat with enough sample behind it into one concrete adjustment.

Read first: Recognising player types · Advice, not a lock: every lesson opens.

Concepts used: GTO and exploitative play, EV, EV loss and how Rail grades

What each number counts

Every stat in the opponents panel is a count over chances, and the chances are not the hands. Putting money in before the flop has a chance every hand; a 3-bet only when someone opened first; folding to a flop c-bet only when he called preflop and then faced the raiser's bet on the flop. A player you have 300 hands on may have faced a c-bet a few dozen times.

• Put money in and raised before the flop: a chance every hand dealt.
• 3-bet: a chance each time someone opened before him.
• Folded to a 3-bet: a chance each time he raised and was 3-bet.
• C-bet the flop: a chance each time he raised preflop and saw the flop.
• Folded to a flop c-bet: a chance each time he faced one.
• Went to showdown and won at showdown: chances when he saw the flop, and when he reached showdown.
• Bet or raised after the flop: a chance at every postflop decision.

How wide the truth can be

A stat seen p of the time over n chances can be off from the player's real frequency by about 1.96 × √(p(1 − p) / n) either way, 95 times in 100. At 30% over 50 chances that is ± 12.7 points: anything from about 17% to 43% fits. Over 500 chances it is ± 4.0.

To get within ± 5 points a stat near 30% needs about 322.7 chances, call it 323. Halving the margin to ± 2.5 needs 1290.8: four times as many. Stats near the edges settle faster: 10% over 50 chances is already ± 8.3.

From a stat to a decision

A stat earns an adjustment when its whole interval sits on one side of a line that decides something. Folds to a c-bet decide whether a pure bluff of a given size profits: a half-pot bluff needs 33.3% folds. A player at 55% ± 8 is above that line; one at 40% ± 14 is not, whatever the 40% suggests.

Stats that come every hand settle in a few hundred hands. Stats that need a rarer spot — folding to a 3-bet, anything on the river — need many sessions, and on one player they may never settle. The pool section below sums your own opponents, which is where rarer stats first become readable.

Rules of thumb

• Count chances, not hands.
• Margin ≈ 1.96 × √(p(1 − p) / n); four times the chances for half the margin.
• Act on a stat when its whole interval clears the line that decides.
• Rare spots need your whole pool, not one player.

When they break

• A stat from a different game or stake describes a different player.
• Players change; a big old sample can hide a recent change.

Your own pool

Practice

The lesson is mastered when every exercise below that Rail can grade is passed. Missed items come back later as review cards.