This dice probability calculator computes the exact odds for rolling any sum with any number of six-sided dice — P(exactly the sum), P(at least), P(at most) — from the full outcome distribution, plus the expected value of the roll.

–

Counting Every Way the Cubes Can Fall

Dice probability is the founding exercise of the whole subject — Pascal and Fermat invented probability theory in 1654 to answer a gambler’s dice questions — and its method is pure counting: enumerate the equally likely outcomes, count the favorable ones, divide. Two dice have 36 ordered outcomes; the sum 7 is assembled six ways and 2 only one way, which is the entire story of the triangular 2d6 distribution and of craps pricing. This calculator generalizes the count by convolution — folding one die’s distribution into the running total at a time — so any number of dice up to 20 and any target sum gets exact counts, not simulations: P(7 with 2d6) = 6/36 = 16.67%, P(10 with 3d6) = 27/216 = 12.5%, and the expected sum is always 3.5 × the number of dice. Watch the shape as n grows: flat (1 die) → triangular (2) → visibly Gaussian (4+) — the central limit theorem, the deepest fact in statistics, assembling itself on a gaming table.

The distinctions that separate exact answers from bar arguments: outcomes are ORDERED (a 1-then-6 and a 6-then-1 are two outcomes, which is why 7 has six ways and not three); “at least” questions are cumulative sums over the distribution (P(≥10 on 2d6) = (3+2+1)/36 = 16.67% — sums 10, 11 and 12 carry 3, 2 and 1 ways, mirroring P(≤4) by the distribution’s symmetry); and specific-face questions (“at least one six in four rolls”) are NOT sum questions — they belong to the binomial/complement family: 1 − (5/6)⁴ = 51.77%, the very computation that started the Pascal–Fermat correspondence when the Chevalier de Méré noticed his two dice bets disagreed with his arithmetic. This page prices sums exactly; the complement rule next door prices appearances.

Worked Examples (Verified)

  • Defaults: 2d6, sum 7 → 16.67% (6 of 36); P(≤7) = 58.33%, P(≥7) = 58.33% — the median straddles the mode.
  • Snake eyes & boxcars: P(2) = P(12) = 2.78% each (1 of 36).
  • 3d6: sum 10 or 11 → 12.5% each (27 of 216) — the twin peaks of the 3d6 bell.
  • 4d6: sum 14 (the mean) → 11.27% (146 of 1,296).
  • 1d6: every face 16.67% — the flat ancestor of all the curves above.

Dice Probability Discipline

  • Ordered outcomes: count (1,6) and (6,1) separately — halving them is the most common dice error ever made.
  • Sums vs faces are different distributions: P(sum ≥ 10) and P(at least one 6) live in different calculators for a reason.
  • Expected value 3.5n is a long-run average, not a prediction: no single roll of 2d6 can equal the 7.0 average’s fractional cousins — 7 itself is merely the mode.
  • Independence is assumed (fair dice, no memory): loaded dice and worn casino dice are empirical questions, answered with chi-square tests on counts, not this formula.
  • Advantage/disadvantage mechanics (roll 2, take higher) reshape the distribution entirely — the maximum of two dice has P(max = 6) = 11/36 = 30.56%, computed by CDF squaring, not by the sum convolution here.

Related Tools

Frequently Asked Questions

What is the probability of rolling a 7 with two dice?

16.67% — 6 of the 36 equally likely ordered outcomes (1-6, 2-5, 3-4, 4-3, 5-2, 6-1) sum to 7, making it the single most likely sum in 2d6. The full distribution is triangular: 2 and 12 at 2.78% each, rising by 2.78 points per step to the 7 peak. This calculator produces the exact distribution for any number of dice and any target sum, with the cumulative “at least/at most” odds alongside.

Why is 7 the most common roll in craps?

Combinatorics: seven is the sum with the most ordered decompositions across two dice (six ways), sitting at the center of the 2–12 range with maximum pairing freedom. Craps is engineered on this asymmetry — the come-out 7 wins for pass bettors, the post-point 7 (“seven out”) is the shooter’s doom, and every proposition bet on the layout is a wager about this exact distribution’s shape, priced with a house edge computed from the same counts this calculator enumerates.

What is the expected value of a dice roll?

3.5 per die — the average of faces 1–6 — so a roll of n dice expects 3.5n in total (7.0 for 2d6, 10.5 for 3d6). It’s a long-run mean: the single roll 2d6 can’t show 7.0’s decimal relatives, but average 10,000 rolls and the mean homes to 7.00 within a whisker (the standard error shrinks with √rolls — the law of large numbers, dice edition). Wagers priced off 3.5n versus offered payouts are how table-game edges are computed.

How do you calculate the probability of rolling at least a certain sum?

Add the probabilities of that sum and every sum above it — the cumulative tail of the distribution. P(≥10 with 2d6) = P(10) + P(11) + P(12) = (3 + 2 + 1)/36 = 16.67%. This calculator reports P(≥ target) and P(≤ target) with every exact probability, computed by full enumeration (convolution), so the tails are exact for up to 20 dice — the same arithmetic strategy guides use for RPG damage curves and board-game movement odds.

What is the probability of rolling at least one 6 in four rolls?

Use the complement, not the sum distribution: P(no six in 4 rolls) = (5/6)⁴ = 48.23%, so P(at least one six) = 51.77%. This was literally the founding problem of probability — the Chevalier de Méré’s observation that this bet is profitable while the analogous “at least one double-six in 24 rolls of two dice” (1 − (35/36)²⁴ = 49.14%) is not, despite his scaling intuition saying both should match. Face-appearance questions belong to the binomial/complement family; sum questions belong to this calculator.

How does the distribution change with more dice?

It bells. One die: flat. Two: a perfect triangle. Three: a smooth hump (10 and 11 tie at 12.5%). Four and beyond: indistinguishable from a normal curve by eye, with mean 3.5n and standard deviation 1.71×√n — the central limit theorem in its friendliest demonstration, because each die is an independent uniform contribution. It’s also why RPG systems chose 3d6 for ability scores: the bell concentrates characters near average with rare extremes, precisely unlike the flat spread of a single d20.

Are online dice probability calculators using simulation?

The good ones don’t need to — dice counts are small enough for exact enumeration. This calculator convolves the count distribution exactly (integers, no randomness), so its percentages are the true values to two decimals, not Monte Carlo estimates that wobble between runs. Simulation earns its keep when enumeration explodes (dice pools with exploding/reroll mechanics, dozens of dice with conditional rules); for plain sums up to 20 dice, counting beats sampling on both speed and certainty.

Is this dice probability calculator free?

Yes — free, instant, no sign-up. Enter the number of dice and a target sum for the exact probability, both cumulative tails, outcome counts and the expected sum.