Dice Roller

The six most common dice shapes

Most tabletop games draw from a small family of dice shapes, each one a Platonic or Catalan solid chosen because every one of its faces is identical in size and shape, which is what makes a fair roll possible in the first place. Here is what the Non-Conventional Dice Roller's quick-pick buttons actually represent:

• d4 (tetrahedron, 4 faces). A triangular pyramid with four triangular faces. Unlike every other die on this list, a d4 has no single face that lands “up”; instead, the result is read from whichever vertex points upward, or from the number printed along the bottom edge of the triangle facing the roller.
• d6 (cube, 6 faces), also called 6 sided dice. The familiar cube used in board games worldwide, six square faces, each showing a number from 1 to 6, usually as pips rather than digits. It is the oldest and most common die shape by a wide margin, and the shape the “Roll Two Dice” tool above uses.
• d8 (octahedron, 8 faces). Two square pyramids joined at their square base, giving eight triangular faces. Common in tabletop role-playing systems for weapon damage rolls that need a wider spread than a d6 without jumping straight to a d10.
• d10 (pentagonal trapezohedron, 10 faces). Ten kite-shaped faces arranged around a central axis, numbered 0 through 9. Two d10s rolled together and read as a two-digit number (with 0 standing in for 10 in the tens place) produce a percentile roll, a number from 1 to 100.
• d12 (dodecahedron, 12 faces). Twelve regular pentagonal faces. Less common than the d6, d8, d10 or d20, but a fixture in certain tabletop systems for specific weapon or spell effects that call for a slightly wider range than a d10.
• d20 (icosahedron, 20 faces). Twenty triangular faces, and the signature die of modern tabletop role-playing games, where it usually decides whether an action succeeds or fails. Its 20-sided spread gives a wide, evenly distributed range for a single roll, which is exactly why the Non-Conventional Dice Roller's d20 quick-pick button is one of its most used.

A “percentile die,” or d100, is not usually its own physical shape. It is almost always simulated with a pair of d10s (one read as the tens digit, one as the ones digit) or, in a digital tool like this one, by simply setting sides per die to 100 directly and rolling a single virtual die across that full range.

How the random roll works

Each die in a roll is drawn independently, the same way a physical die lands on one face without any influence from the die next to it. Behind the scenes, the calculator generates a random fraction between 0 and 1, scales it up to the number of sides you chose, and rounds down to land on one whole number inside that range. That process spreads the result evenly across every face: with a six-sided die, the chance of landing on 1 is the same as the chance of landing on 6, and every number in between, because each face covers an equal slice of the 0 to 1 range the random draw starts from.

This is what people mean when they call a die roll a uniform distribution. No face is weighted more heavily than another, and the outcome of one die has no effect on the outcome of the next one you roll, whether that next die is part of the same throw or a completely separate roll later on. Roll a d20 twenty times and you should expect roughly one appearance of each number over the long run, though any short run of rolls, like any short run of a real die, can look streaky purely by chance.

The underlying random source is JavaScript's standard Math.random(), the same pseudorandom generator behind the random number generator on this site. It is fast and more than good enough for games, teaching, and everyday decisions, but it is not built to resist someone deliberately trying to predict it, so it should never be used anywhere a genuinely unguessable outcome matters, such as security codes or cryptographic keys.

Common uses for a dice roller

A digital dice roller covers most of the situations a physical die would, without needing to find one, and adds a few uses a handful of physical dice cannot cover as conveniently:

• Tabletop and board games. Roll the exact dice a game calls for, from a single d6 for a classic board game to a handful of d20s for a tabletop role-playing session, without hunting down a matching physical set.
• Decision-making. Use a simple d2 or d6 roll to break a tie, choose who goes first, or pick between a short list of options at random, the same way flipping a coin or drawing straws would.
• Teaching probability. Rolling the same die repeatedly and watching how the results spread out over many throws is a direct, hands-on way to see a uniform distribution in action, and to compare a small run of results against the long-run expectation.
• Simulating a specific dice pool. Many games and systems call for rolling several dice of the same type together and summing them, such as two six-sided dice for a classic 2d6 roll; setting the number of dice and sides once reproduces exactly that pool on every press of Roll.

How random is a physical dice, really?

It is tempting to assume a physical die tumbling across a table is the “most” random option available, and a computer-generated roll is somehow a lesser substitute. In practice the comparison runs the other way more often than people expect. A real die is a manufactured object, and manufacturing variance, a fraction of a millimetre of extra material around the drilled pips on one face, a slightly rounded edge from a worn mold, a marginally off-center weight distribution, can nudge a physical die's odds away from a perfectly even one-in-six per face, especially over a very large number of rolls. Casino-grade dice go through strict manufacturing tolerances specifically to minimize this; a cheap board-game die usually has not.

A virtual die has no manufacturing tolerance to worry about at all. Each roll here comes from scaling a fresh pseudorandom fraction to the number of sides you asked for (see the next section for exactly how), which spreads results evenly across every face by construction, not by physical precision. That does not make it flawless: a pseudorandom generator is deterministic under the hood, not truly random in the deepest sense, and it is not built to resist someone deliberately trying to predict it. But for the everyday purpose of rolling dice fairly, for a game, a classroom demonstration, or breaking a tie, a well-built virtual roll is, if anything, closer to the theoretical ideal of a uniform distribution than most physical dice actually achieve in practice.

Reading the results: individual rolls versus the total

Every roll shows two things at once, and which one matters depends on what you are using it for. The row of individual results is what most tabletop games actually care about, since rules often key off a specific die's face rather than the combined sum, such as a game where only one particular die in the pool determines a critical outcome. The total underneath adds every die together, which is what a game calling for a combined score, like a 2d6 damage roll or a percentile check, is actually looking for.

Because both numbers are shown together, you do not need to add up the dice yourself or re-roll to check one or the other. If your use case genuinely only needs a single random whole number rather than a set of dice results, the random number generator covers that directly with a custom min and max range. And if you are working out the odds behind a particular roll, such as the chance of rolling a specific total on two dice, the probability calculator can work through those odds separately.