LearnModule X1 · Reading people
X1-L1
Recognising player types
How to place a player by how many hands he plays and how aggressively, and the one adjustment Rail's exploit lab finds against each type.
In this lesson
- Place a player by two questions: how many hands, and how aggressively.
- Read a type from the opponents panel, with the sample behind it.
- Name the main adjustment against folders, callers, passive players and bluffers.
- Treat a type as a read that can be wrong, and fall back to the baseline.
Two questions about a player
Every label players use for each other answers two questions. How many hands does this player play? And when he plays one, does he bet and raise, or check and call? Tight or loose is the first answer, passive or aggressive the second, and the four corners are the familiar types: tight and passive, loose and passive (the calling station), tight and aggressive, loose and aggressive (the maniac, at the far end).
Rail's opponents panel answers both from your own hands: how often a player puts money in and raises before the flop for the first question, and how often he bets or raises after the flop for the second. A number is only as good as the chances behind it. A player who played 40% of 100 hands could just as well be a 30% or a 50% player: the 95% interval reaches about 9.6 points either way. The next lesson does that arithmetic properly.
- 95% interval
- ± 9.6%
- Plausible range
- 30.4% – 49.6%
- Chances for ± 5 points
- 369
Each type has one main adjustment
A type is useful only if it changes what you do. Rail's exploit lab locks one tendency of the opponent on a river it solves and computes the best response; over many rivers the adjustments it finds are these:
- A player who folds too much to bets: bluff with far more of the hands that cannot win at showdown, turn weak pairs and ace-high into bluffs, and bet your strongest hands smaller.
- A player who calls too much: stop bluffing, bet your strong hands bigger, and bet some middle pairs you would check against the solve.
- A player who never raises: bet thinner, since nothing punishes a thin bet.
- A player who bluffs too little: fold more of your bluff-catchers.
- A player who bluffs too much: call down wider with your pairs and ace-high.
A type is a read, not a fact
Putting a player in a corner is a guess about how he plays the spots you have not seen yet. Each adjustment above gives something back if the guess is wrong, and the lab measures how much (lesson X3-L1). A label made on a thin sample, or one a player has outgrown, is worse than none: the baseline is what you play while the read is unclear.
Rules of thumb
- Ask two questions: how many hands, and how aggressively.
- Turn a type into one adjustment, or it is only a label.
- Against folders, bluff more; against callers, value bet more and bluff less.
- Read the sample before the number.
When they break
- A player can be loose before the flop and tight after it: type him by the spot you are in.
- Players change during a session; a label from last month may be wrong tonight.
Example hands
Examples come from two places only: your own analysed hands in this lesson's spots, and a hand Rail deals and grades itself. Never a hand from outside Rail.
Rail scripts this hand from the lesson's own practice, always the same one, and grades it with the analysis. Decide first, then read Rail's verdict.
Your own pool
Your opponents' tendencies from your own hands, summed over the opponents panel, with how many chances each stat had and the 95% interval that sample allows.
Practice
The lesson is mastered when every exercise below that Rail can grade is passed. Missed items come back later as review cards.
Exercise 1
Pass: 2 of 4 rightFour rivers, each with one of the five reads locked: play the hand Rail's best response plays against it.
Exercise 2
Pass: 5 of 6 rightSix opponents' folds to a flop c-bet, each over a number of chances: say what the stat supports — folding more than a half-pot bluff needs, less than a third-pot bluff needs, or no read yet — by the same rule your own pool section uses.