The shortcut is useful only after the counting is sound. “I have a straight draw and a flush draw” does not automatically mean 17 outs. “The flush gets there” does not automatically mean you win. And a two-card chance does not mean a single flop call pays for both cards. This guide separates those questions before putting a percentage beside your hand.
What Counts as an Out?
Start by naming the result you need. A card can improve your hand category without beating the opponent. If you have ace-high and the opponent has a set, an ace gives you a pair but not the best hand. If you have a small flush draw against a higher flush draw, completing the flush can still leave you behind. Count cards toward a defined outcome, not toward any improvement whatsoever.
In an original open-ended example, Hero holds 8♥7♥ on 6♥5♣K♥. Any 4 or 9 completes a straight on the next card: four fours plus four nines, eight cards before additional known-card restrictions. A gutshot is narrower. With 8♦7♣ on J♠T♦2♣, only a 9 completes the straight on the next card, so there are four rank-based straight-completion cards.
For a flush draw, count the suit. Hero's two hearts plus the flop's two hearts account for four of the thirteen hearts, leaving nine heart cards unseen. That is a count of cards that complete the flush, not an assertion that each is a clean winning out against every range. The hand rankings guide explains which made hands can still beat it.
Count the Same Card Only Once
Return to 8♥7♥ on 6♥5♣K♥. The flush-completion set has nine hearts. The straight-completion set has the four fours and four nines. But 4♥ and 9♥ belong to both sets. They are each one physical card, not two independent chances. The union contains 9 + 8 − 2 = 15 distinct cards.
Write the list when a combo draw is unfamiliar. Hearts: A♥, Q♥, J♥, T♥, 9♥, 5♥, 4♥, 3♥, 2♥. Additional straight cards: 4♣, 4♦, 4♠, 9♣, 9♦, 9♠. The list has fifteen entries, contains no known card, and has no duplicates. A simple list catches more mistakes than assigning an impressive-sounding draw label.
Overcards need the same discipline. An ace or king may be a useful improvement against one pair, but not against every two-pair, set, or straight. Do not add six automatic overcard outs to a flush draw simply because two high ranks are in your hand. First ask which opposing hands those pairs beat and which exact cards are already included in the flush count.
Clean Outs, Dirty Outs, and Redraws
A clean out wins in the relevant model when it arrives. A dirty out appears to improve you but can improve an opponent more, complete a better draw, or create a tie rather than a win. Cleanliness depends on the opponent's hand or range and on whether another card remains. The same card can be clean against top pair and dirty against a set.
Consider Hero A♥J♥ on 9♥6♥2♣ against the now-known 9♣9♦. The 2♥ completes Hero's flush, but it also pairs the board. The opponent then has nines full of twos and beats the flush. Of the nine unseen hearts, eight put Hero ahead on the turn; 2♥ does not. With both hands exposed on this flop, seven cards are known, so 45 turn cards remain, not 47.
Even a turn that puts you ahead may not lock up the pot. On a safe heart turn in that example, the set can still improve to a full house or quads on some rivers. That is a redraw: a later improvement that can beat the hand you just made. Counting the chance to lead on the next card and counting final showdown equity are different calculations.
When the opponent's cards are unknown, a fixed discount such as “half an out for every overcard” is only a heuristic. A more transparent review compares explicit ranges: one in which the pair cards win often, another in which they rarely do. If the decision changes between them, record that range uncertainty. Do not present a fractional-out shortcut as an exact card count.
One Card to Come vs. Two Cards to Come
The PokerBank's percentage odds chart separates flop-to-turn, turn-to-river, and flop-to-river probabilities. Its usage notes also distinguish seeing one card from an all-in situation in which no further call is needed. A two-card percentage cannot be substituted for the next-card chance merely because the hand is currently on the flop.
With only Hero's two cards and the three-card flop known, 52 − 2 − 3 = 47 cards are unseen. For a fixed set of o qualifying cards, the next-card probability is o/47. On the turn, with Hero's two cards and four board cards known, 46 are unseen, so the next-card probability is o/46. These are information models: an opponent's unexposed hole cards have not been arbitrarily removed from the calculation.
If both players' hands are known, the denominators change. A heads-up flop with both hands exposed leaves 45 possible turn cards, followed by 44 possible rivers. On the turn with both hands exposed, there are 44 possible rivers. Do not use a television all-in display to “verify” a 47/46 table unless the information sets and winning-out conditions actually match.
For two cards after the flop and a fixed qualifying set, calculate the chance of missing twice and subtract it from 1: p = 1 − ((47 − o)/47) × ((46 − o)/46). The second miss numerator is one smaller because the first missed card is gone. This is the chance of seeing at least one card from that original set, not a full hand evaluator that detects new outs, redraws, or board-created ties.
The Rule of 4 and 2
The conventional rule multiplies a fixed out count by 4 for two future cards and by 2 for one. It is an estimate, not an exact formula. Upswing explicitly notes that the two-card estimate becomes less accurate as the out count grows. Use an exact calculation when a small difference could change the review's conclusion.
For eight outs, the rule gives 16% on the next card and 32% over two cards. The flop-to-turn exact chance is 8/47 = 17.02%; the turn-to-river exact chance is 8/46 = 17.39%; the fixed-set two-card chance is 31.45%. The shortcut is not biased in one direction for every horizon: it understates these one-card figures but slightly overstates the two-card figure.
Keep percentage points separate from percent change. An estimate of 60% instead of 54.12% is high by 5.88 percentage points. It is not high by only “5.88%” relative to the true value. For a decision near its threshold, reporting the absolute percentage-point error is the clearest comparison.
Exact Probabilities and Approximation Error
This original table uses the 47-card flop information model and a constant qualifying set. “One Card” means the next turn card, not the river after a known turn. Errors are absolute differences in percentage points, rounded to two decimals after calculation. The two-card column is valid for the event “at least one of these o cards appears”; it is not automatically equity.
Duddhawork derives the two-card probability and compares it with the rule of 4 and 2 in English, with English captions. This is the chance of completing the specified draw, not necessarily winning; dirty outs and further payments still matter.
| Outs | One Card Exact, o/47 | 2o Estimate | Absolute Error, pp | Two Cards Exact | 4o Estimate | Absolute Error, pp |
|---|---|---|---|---|---|---|
| 1 | 2.13% | 2.00% | 0.13 | 4.26% | 4.00% | 0.26 |
| 2 | 4.26% | 4.00% | 0.26 | 8.42% | 8.00% | 0.42 |
| 3 | 6.38% | 6.00% | 0.38 | 12.49% | 12.00% | 0.49 |
| 4 | 8.51% | 8.00% | 0.51 | 16.47% | 16.00% | 0.47 |
| 5 | 10.64% | 10.00% | 0.64 | 20.35% | 20.00% | 0.35 |
| 6 | 12.77% | 12.00% | 0.77 | 24.14% | 24.00% | 0.14 |
| 7 | 14.89% | 14.00% | 0.89 | 27.84% | 28.00% | 0.16 |
| 8 | 17.02% | 16.00% | 1.02 | 31.45% | 32.00% | 0.55 |
| 9 | 19.15% | 18.00% | 1.15 | 34.97% | 36.00% | 1.03 |
| 10 | 21.28% | 20.00% | 1.28 | 38.39% | 40.00% | 1.61 |
| 11 | 23.40% | 22.00% | 1.40 | 41.72% | 44.00% | 2.28 |
| 12 | 25.53% | 24.00% | 1.53 | 44.96% | 48.00% | 3.04 |
| 13 | 27.66% | 26.00% | 1.66 | 48.10% | 52.00% | 3.90 |
| 14 | 29.79% | 28.00% | 1.79 | 51.16% | 56.00% | 4.84 |
| 15 | 31.91% | 30.00% | 1.91 | 54.12% | 60.00% | 5.88 |
| 16 | 34.04% | 32.00% | 2.04 | 56.98% | 64.00% | 7.02 |
| 17 | 36.17% | 34.00% | 2.17 | 59.76% | 68.00% | 8.24 |
| 18 | 38.30% | 36.00% | 2.30 | 62.44% | 72.00% | 9.56 |
| 19 | 40.43% | 38.00% | 2.43 | 65.03% | 76.00% | 10.97 |
| 20 | 42.55% | 40.00% | 2.55 | 67.53% | 80.00% | 12.47 |
To reproduce the nine-out row: one card = 9/47 = 19.15%. Two cards = 1 − (38/47 × 37/46) = 34.97%. The estimates are 18% and 36%, giving absolute errors of 1.15 and 1.03 percentage points. The same procedure generates every row; no chart value is a measured player win rate.
Combo Draws and Backdoor Draws
The fifteen-card union in the earlier example produces a 54.12% chance to see at least one of those original cards over two future streets in the 47/46 model. It does not prove a 54.12% chance to win. The opponent can hold blockers, make a higher hand, or redraw after you improve. Other runouts can also help Hero without containing one of the initial fifteen cards. Use full hand-versus-range enumeration when showdown equity is the question.
A backdoor draw needs help on both remaining streets. Hero A♥J♥ on 9♥6♣2♦ has three hearts among the five known cards and needs two more hearts for a flush. Ten hearts remain unseen. The flush-completion probability is (10/47) × (9/46) = 4.16%, assuming no other known cards. That is not a one-card “ten-out flush draw.” Most turns fail to preserve the backdoor flush possibility.
You can check that result with unordered pairs: choose any two of ten hearts, giving 45 pairs, out of 47 choose 2 = 1,081 total two-card runouts. Thus 45/1,081 = 4.16%. Both methods describe the same event. Neither checks whether a completed flush beats the opponent, and neither says how much it costs to see the river.
A Worked Hand and Answer Key
Try this turn first: Hero 8♥7♥, board 6♥5♣K♥2♠, opponent's hand unknown. List the straight and flush completion cards. There are still nine hearts and eight fours/nines, with two overlaps, so fifteen distinct improvement cards. Six cards are known, leaving 46 unseen. The next-card completion chance is 15/46 = 32.61%; the rule of 2 gives 30%.
Now add a condition: the opponent shows A♠A♣ and is all-in. There are eight known cards, so 44 rivers remain. Neither ace blocks the fifteen completion cards. Every one of the fifteen makes Hero a straight or flush that beats one pair, and no remaining non-completion card gives Hero a winning hand. There are no ties. Here Hero's exact equity is 15/44 = 34.09%. The answer changed because the information model changed, not because the shortcut improved.
| Exercise | Correct Count or Result | Why |
|---|---|---|
| 8♥7♥ on 6♥5♣K♥, count straight-or-flush completion cards | 15 distinct cards | Two cards belong to both draws |
| 8♦7♣ on J♠T♦2♣, count immediate straight cards | Four nines | This is a gutshot, not an open-ender |
| A♥J♥ on 9♥6♥2♣ versus 9♣9♦: is 2♥ a clean turn out? | No | The opponent makes a full house |
| Nine fixed outs, flop to turn, opponent unknown | 9/47 = 19.15% | Five cards are known |
| Nine fixed outs, turn to river, opponent unknown | 9/46 = 19.57% | Six cards are known |
| Backdoor hearts: A♥J♥ on 9♥6♣2♦ | 45/1,081 = 4.16% to complete a flush | Two hearts must arrive |
| Worked combo draw on the turn against exposed A♠A♣ | 15/44 = 34.09% equity | Eight known cards; all fifteen specified rivers win |
When you review a draw, save the exact cards and three separate statements: the counted event, its probability, and the price of continuing. For example: “Fifteen rivers make a straight or flush; under the unknown-hand model that is 32.61%; profitability still needs the opponent's range and betting model.” This wording prevents a completion percentage from silently becoming a recommendation to call.
Use pot odds to calculate the call's price and poker hand probabilities for broader hand-frequency questions. The useful habit is not memorizing every table cell. It is counting each physical card once, checking what it actually beats, and choosing a probability model that matches the cards you will see.
Sources and Calculation Notes
- Upswing Poker — “Poker Drawing Odds & Outs Explained” (October 4, 2024), “What Are Outs in Poker?”, “Combo Draws,” and “The Rule of 4 and 2.” Supports the terminology and conventional approximation. All example hands here are original and independently counted; the source's open-ended example is not reused. Checked October 28, 2026.
- The PokerBank — “Poker Percentage Odds Chart”, “Percentage Odds Chart” and “How to Use the Percentage Odds Chart.” Supports distinguishing next-card from two-card horizons. No publisher date was visible. Checked October 28, 2026.
The 1–20-out table, approximation errors, card lists, exposed-hand comparisons, backdoor calculation, and exercises are original calculations from the stated 52-card models. The table assumes a fixed qualifying set; it is not a substitute for exact showdown evaluation when outs change or an opponent can redraw.
About the Editorial Team
The Poker-Eye Editorial Team researches, checks, and maintains product and educational content. Sources and product boundaries are reviewed before publication.