In chapter 24 of The Theory of Poker, David Sklansky focuses on how to analyze poker decisions in real time at the table, using risk–reward thinking and structured reasoning, even when you only have a few seconds.
Poker as Risk vs. Reward
Sklansky frames every poker decision as:
-
Risk: The bet(s) you might lose.
-
Reward:
-
The current pot, plus
-
Any future benefits (e.g., advertising value of a bluff, future action you might earn, or extra bets later).
-
Examples of this lens:
-
Bluffing:
-
Risk = current bet.
-
Reward = pot + the image you create.
-
-
Betting a medium-strength hand before all cards are out:
-
Risk = your bet.
-
Reward = denying a free card that could beat you.
-
-
Checking a big hand:
-
Risk = missing a bet now and occasionally losing to a hand that would have folded.
-
Reward = bigger payoffs later through check-raises or extra calls.
-
Good players constantly (and quickly) ask:
-
What do I gain by this action (now and later)?
-
What can it cost me?
-
Is the gain worth the risk?
The Time Pressure Problem
In theory you could sit and compute odds and casework on every hand, but:
-
Poker is fast; you usually only have seconds.
-
Long pauses give away that you’re facing a tough decision.
-
Strong players tend to combine:
-
Quick thinking and
-
Instant recall of past mistakes (their own and others’).
-
So the ideas in this chapter are about building mental habits that let you approximate good analysis almost automatically.
How Strong Players Analyze Hands in Theory
Sklansky contrasts two mental approaches:
1. Weak Approach: Locking in One Hand
Many players:
-
Decide “He has this hand,” then
-
Choose their action as if that were certainly true.
This is dangerous, because it ignores the full range of possible holdings and leads to big, repeated errors.
2. Strong Approach: Ranges and Probabilities
Better players:
-
List the possible hands the opponent could have.
-
Estimate the likelihood of each hand.
-
Identify the best play against each candidate hand.
-
Choose the action that is best most often, given those probabilities.
Sometimes all lines lead to the same action (e.g., easy folds), but often:
-
Different opponent hands would dictate different plays.
-
You then choose the play that’s right against the most likely hand(s) or against the weight of the range (e.g., right 60% of the time vs. 40%).
He illustrates this with examples where:
-
Bluff-raise is good if villain has a specific weak hand,
-
But terrible if villain has his other likely holdings (like two pair or a straight),
-
So you only attempt it when the weak-hand case is both likely enough and villain is the type who can fold.
Example: Draw Poker with Three Aces
You raise with three aces in draw, get reraised by a very tight player who has either:
-
Trips (smaller than your aces), or
-
A made hand (e.g., straight/flush).
Based on card distribution, he has:
-
Trips about 65% of the time,
-
A pat hand about 35% of the time.
Instead of reraising pre-draw, Sklansky shows:
-
It’s better to just call, draw one, and then:
-
Check–raise when he draws (indicating trips),
-
Or save money when he stands pat (indicating a made hand).
-
That line:
-
Wins more when you’re ahead,
-
Loses less when you’re behind,
-
And is therefore correct regardless of which hand he actually has.
This is a clean example of:
-
Considering multiple possibilities,
-
Seeing that one line dominates both cases.
Example: Hold’em with Top Pair in a Small Pot
In a limit hold’em hand:
-
You have K K on a dry board.
-
Villain calls the flop, then checks the turn.
-
You assign him a mix of:
-
Better hands (slowplays),
-
Decent kings,
-
Mediocre hands,
-
Straight draws.
-
Sklansky’s reasoning:
-
If you bet turn, villain:
-
Calls with his decent hands and draws,
-
Folds many marginal hands that you currently beat.
-
-
If you check turn:
-
You give a free card (risk),
-
But you:
-
Induce calls on the river from hands that would have folded the turn,
-
Sometimes induce bluffs on the river.
-
-
In a small pot, the extra future value from weaker calls and bluffs can outweigh the risk of the free card, so checking behind on the turn may be best.
This shows:
-
How your action can shape villain’s calling/bluffing range,
-
And how that factors into risk–reward.
When the “Less Frequent” Play Is Actually Better
A key twist in this chapter:
Even if a play is theoretically correct more often than another, it may still be worse overall if its mistakes are much more expensive.
Calling vs. Folding: Cost of Errors
-
If you call and you’re wrong:
-
You lose one bet.
-
-
If you fold the best hand:
-
You lose the entire pot.
-
So even if you think you’re behind more than half of the time, the risk–reward ratio can still favor calling, depending on pot odds.
The general idea:
-
The severity of a mistake matters as much as its frequency.
Example: Queens vs. a Preflop Raise (No-Limit)
You have QQ; villain raises with a range you know is only:
-
AA, KK, or AK.
Math says he has AK more often than AA/KK (4:3). Lifetimes:
-
You’re a small favorite versus AK,
-
But a huge dog versus AA or KK.
So:
-
Even though you’re ahead on average vs. his range, the times you’re crushed by AA/KK cost so much that simply calling is not automatically correct.
-
Sometimes folding or taking a different line (depending on stack sizes and game context) is better.
This highlights how expected value depends on both:
-
Probability of being right/wrong
-
And the size of the swing when you are.
Example: Razz – Raising Even When You’re Often Behind
In a razz hand:
-
You have a 9-low draw against an opponent whose board suggests either:
-
A made strong low (slightly more likely), or
-
A paired board (you’re actually ahead but he’s representing strength).
-
-
Calling is correct in more cases (by raw frequency).
-
But when you raise:
-
You gain a lot when you’re ahead (he’s often in bad shape).
-
You’re not in terrible shape even when you’re slightly behind.
-
So raising, though “wrong” slightly more often, yields higher long-run expectation because your good scenarios are so much more profitable.
The Meta-Principles of At-the-Table Analysis
Sklansky closes the chapter with a big-picture summary:
-
Think in ranges, not single hands
-
List possible hands, assign rough probabilities, and analyze your options against that mix.
-
-
Compare plays by how often they are right and how bad they are when wrong
-
Don’t just pick the move that’s right more than 50% of the time;
-
Consider whether its mistakes are catastrophic or small.
-
-
Risk–reward is always central
-
Calling vs. folding, betting vs. checking, raising vs. calling all hinge on:
-
Size of the pot,
-
Size of the bet,
-
Likelihood of each opponent hand,
-
And the impact of being wrong.
-
-
-
Speed comes from practice
-
The chapter’s math and logic are not meant for full, explicit calculation every hand.
-
They’re meant to train your intuition so that, under time pressure, your “snap decisions” tend to align with good theory.
-
Final Takeaway
Chapter 24 is about turning all the earlier concepts — pot odds, hand reading, bluffing, psychology, game theory — into fast, practical decision-making at the table. You don’t just ask “Am I ahead?”; you ask:
-
What could he have?
-
How often?
-
What happens if I’m wrong this way vs. that way?
-
Is this risk acceptable for this reward?
The more naturally you can run that mental script in real time, the closer you get to playing like a top-level professional.