Range Thinking in Poker: Beyond Reading Single Hands

Think about possible hands, not one guessed hand

A range is a set of possible two-card combinations with a weight for each one in a particular situation. You do not see an opponent’s private cards, so “they have A-K” is a hypothesis, not a fact. Their position, earlier actions, the board and known cards help you judge which possibilities remain and how plausible each is.

A range is only as useful as its assumptions. It cannot make a decision automatically profitable, and a guessed distribution is not the same as observed opponent data or a solver output. The aim here is to make the labels, counts and updates explicit enough to question.

Before the deal, a 52-card deck offers 1,326 distinct two-card combinations. After cards appear, some become impossible. The card-removal example shows the arithmetic.

Read the 13×13 starting-hand matrix

The grid groups physical combinations into 169 labels. Diagonal labels are pairs such as AA and JJ. Above the diagonal, AKs means ace-king of the same suit. Below it, AKo means ace-king of different suits. “s” and “o” describe suits, not how strong the hand is or whether to enter a pot.

A pair has 6 suit combinations, a suited unpaired class has 4, and an offsuit unpaired class has 12. Thus 13 × 6 + 78 × 4 + 78 × 12 = 1,326. The matrix has every rank from ace through deuce, including jack. It is a reference for reading hand labels, not an opening range or a recommendation.

Position changes the question, not the notation

An early-position opener faces more players who can still act. A button opener usually has fewer players left and often acts later after the flop. That can change which hands fit an opening, calling or re-raising range. Stack depth, blinds or antes, rake, earlier action and player tendencies also change the answer; no percentage belongs to a seat in every game.

The small blind is a useful warning against rote charts. Completing or limping can be part of a situation-specific strategy, especially blind-versus-blind or at particular stack depths. “Never limp from the small blind” is not a rule. See the positions guide for seat order, and compare any published range chart with the format it models before using it.

After an action, update weights—do not erase by habit

A call does not prove the player lacks a very strong hand; a check does not prove they missed; and a bet can include different hand strengths. An action changes what is plausible conditional on a model. Only a conflicting known card makes a physical combination impossible.

On the K♠ 9♥ 4♣ flop above, imagine only two alternative exact hands are in a toy model: K♥ Q♥ (top pair) and J♥ T♥ (a gutshot draw). They are alternatives, not cards held simultaneously. Suppose each had a 50% prior weight and, solely for illustration, the player would bet with them 75% and 25% of the time respectively. These are stipulated numbers, not measured behavior or a recommended strategy.

On a narrow screen, focus and scroll the table sideways.

The two unnormalised weights are 0.375 and 0.125; dividing each by their sum, 0.5, gives 75% and 25% within this invented two-hand model. Both hands remain possible after the bet. A different prior or action assumption produces a different result. The turn 7♦ and river 2♠ above extend the same board; the example does not assert what either candidate would do on those streets.

Count combinations still physically available

Before any cards are known, AA has 6 physical combinations: choose two of four aces. On the separate A♠ 7♦ 2♣ flop below, only three aces remain, so AA has 3 combinations. If your known private cards include A♥ as well, only A♦ and A♣ remain; an opponent can hold AA in exactly 1 way. This is card removal, not a claim that they would play that one combo in a given way.

The same principle applies to suited hands. If A♠ is on the board, A♠K♠ cannot be in anyone’s private cards, leaving at most three suited A-K combinations before considering other known cards. When comparing hands, count physical combos and their weights, not just the number of coloured squares in a chart.

Range advantage is not nut advantage

Range advantage compares the weighted average showdown equity of two specified ranges on a specified board. Equity is expected pot share across possible outcomes, including ties—not just the fraction of outright wins. Nut advantage asks which specified range holds more of the strongest available hands. Neither follows automatically from “high board” or “low board” without the ranges, weights and known cards.

To calculate equity for one exact hand against a specified range on a flop, first list each opponent’s surviving physical two-card combination and its weight. For each combination, remove both players’ hole cards and the three board cards; 45 cards remain, giving 45 × 44 ÷ 2 = 990 possible turn-and-river pairs when order does not affect final hand strength. Score both best five-card hands on each runout, count a win as one share and a tie as half a share, then average over runouts and the opponent-combo weights. More players, future betting or an action-dependent range need a different model. The small table above demonstrates weight updating, not showdown equity or whether a bet is profitable.

Practise with an explicit record

1. Write the format, position, effective stacks, action and board before naming a range. Do not borrow a chart from a different situation without its assumptions.
2. List a few candidate physical combinations and remove only cards that conflict with known cards. Distinguish a hand class such as AKs from a particular combo such as A♠K♠.
3. For each action, note why you changed a weight. Keep plausible slow plays, bluffs and checks unless evidence truly excludes them. A single timing tell is weak evidence, not a precise percentage.
4. After the hand, compare your recorded model with the revealed information if there is any. One showdown reveals one hand, not the full range or a reliable population frequency.

What to learn next

• Table positions — who can still act and why seat context matters.
• Hand rankings — evaluate the exact hand, board and kickers before making a range claim.
• GTO basics — game-theory models and their limits.
• Exploitative play — how opponent evidence can change a model.
• Bluffing — value and bluff possibilities without guessing a single hand.
• Poker glossary — definitions for range, combo, blocker and related terms.

Sources and limits

PokerStars’ Hold’em rules state the 169-class/1,326-combination distinction. PokerStars Learn’s combination walkthrough gives the 6/4/12 pair, suited and offsuit counts; the arithmetic here is independently reproducible from a 52-card deck. Its range explanation supports treating hidden hands as possibilities rather than certainty; its sample frequencies are not imported here. Its small-blind discussion illustrates why a universal no-limp rule is unsafe. WPT’s Hold’em rules establish the community-card street order.

The board, blocker and two-hand action examples on this page are deterministic teaching fixtures. Their action weights are hypothetical, not measured opponent frequencies, a solver result, a profitable playbook or proof of an edge.