Probability basics, in plain terms
Every face of a fair die is equally likely: a six-sided die shows 1 as often as 6 over the long run, which is why each face has probability 1 in 6 and why no roll is ever "due" — the die does not remember being kind to you. The most useful number for judging a roll is the average: a d6 averages 3.5 per roll (the faces 1 to 6 add to 21, divided by 6), a d20 averages 10.5, and a d100 averages 50.5. Roll several dice and the total clusters around count × average: two d6 most often total 7, the single most likely outcome, while 2 and 12 are the rare extremes at 1 chance in 36 each. The stats panel shows your roll's average per die next to the minimum and maximum possible, so a lucky or unlucky throw is obvious at a glance.
How the randomness is generated
Where your browser provides it — and modern browsers do — each roll draws from window.crypto.getRandomValues, the cryptographic random source built for security work, mapped onto the die's faces; older browsers fall back to Math.random. Either way every die in a set is rolled independently, exactly like shaking separate physical dice, and nothing about your rolls is stored or sent anywhere: the eight-roll history lives only in this page, for as long as the tab is open. The naming follows tabletop convention: a d20 is a twenty-sided die, a d100 a percentile die, and "2d6" means two six-sided dice added together — the default roll on this page.
The coin flip
Under the dice sits a coin flip with running tallies. A fair coin lands heads half the time over many flips — but "over many" does real work in that sentence: five heads in a row happens by chance in about 1 run of 32, and short streaks are the norm, not evidence of a rigged coin. Flip fifty times with the tally running and the heads share will usually wander towards a half without ever being obliged to arrive exactly. It is the quickest honest demo of how randomness actually behaves, and a perfectly good way to settle who kicks off.
Where dice rolls earn their keep
Board games with lost dice, tabletop role-playing games (where the d20 decides almost everything and the d100 handles percentile skills), classroom probability lessons, raffle-style picks, and everyday decisions that deserve better than dithering — who goes first, which route, which film. For picking a whole number in an arbitrary range rather than a die total, the Random Number Generator is the sibling tool; for shuffling a list of names, use the List Randomizer.
Worked example
Roll 2d6 — two six-sided dice. Suppose they land 3 and 5: the headline total is 8, the average per die is 4.0, and the roll sits inside the possible range of 2 to 12, a little above the expected average total of 7. The same explanation updates with every roll you make: whatever faces appear, the total is simply their sum, and the history table keeps the last eight results so a hot streak can be checked rather than remembered generously.
How to use this calculator
- Choose the die type — d4, d6, d8, d10, d12, d20 or d100 (percentile).
- Set how many dice to roll, from 1 to 6.
- Press Calculate to roll; press again for a fresh roll.
- Read the total in the headline, each die in the table, and the average, minimum and maximum in the stats.
- Scroll to the history to compare your last eight rolls — or use the coin flip below the dice to settle a two-way choice.
Loaded language: fairness and streaks
Two misunderstandings follow dice everywhere. The first is the gambler's fallacy: after five high rolls, a low one feels owed. It is not — independent rolls do not balance themselves in the short run; fairness is a property of thousands of rolls, not the next one. The second is reading meaning into the average of a handful of throws: ten rolls averaging 4.2 on a d6 (expected 3.5) is ordinary luck, well inside normal variation, and casinos and game designers alike rely on players over-reading exactly this kind of noise. Roll for fun, for games and for decisions — and let the history table, not memory, tell you how the dice have actually behaved.