Dice Probability Explained

One die: perfectly flat odds

A six-sided die gives each face a one-in-six chance, about 16.7 percent. No face is favoured over another, which makes it the simplest possible distribution. Statisticians call it uniform.

More faces do not change the character. A D20 gives five percent per face and a D12 about 8.3 percent. As the face count rises, the range of possible values widens and the chance of any specific value drops.

The advantage of a uniform distribution is predictability in design. Every value is equally likely, so the arithmetic stays simple. It suits pass-if-you-roll-at-least checks particularly well.

Why two dice behave differently

The sum of two dice ranges from 2 to 12, but the number of ways to reach each total varies. Seven can be made six ways, from one and six through to six and one. Two can be made exactly one way.

With thirty-six combinations in total, seven comes up six times in thirty-six, about 16.7 percent, while two comes up once, about 2.8 percent. A six-fold difference.

So totals cluster near the middle and thin out towards the extremes, and the effect strengthens as you add dice. It is the most accessible example of what statisticians call the central limit theorem.

• Total of 7: six combinations, about 16.7 percent
• Total of 6 or 8: five each, about 13.9 percent
• Total of 5 or 9: four each, about 11.1 percent
• Total of 2 or 12: one each, about 2.8 percent

Dice count controls volatility

This property is extremely useful in rule design. Two setups can share an average and behave completely differently depending on how many dice you split it across.

For an average around ten, you could roll one D20 or three D6. The first gives every value from 1 to 20 equal weight. The second lands on ten or eleven most of the time and almost never on three or eighteen.

Fewer dice for more luck, more dice for more influence from bonuses and skill. That is precisely why tabletop games roll one D20 for checks and several smaller dice for damage.

Choosing a face count

D4 has a narrow range and suits small modifiers and quick checks. D6 is the most familiar and is ideal for rolling several and approximating a bell curve.

D8 and D12 fill the gap between D6 and D20. D12 is particularly tidy because twelve divides by two, three, four, and six, so results group cleanly. It also works for choosing among twelve options or picking a clock direction.

D20 expresses probability in exact five percent steps, which is why it dominates as a check die. Set a target of twelve and the success rate is precisely forty-five percent.

Using dice to pick options

A common mistake appears when dice are used to choose between options. Rolling a D6 for five options and rerolling on a six is fine. Treating a six as option five is not, because it doubles that option's odds.

Picking among seven options with two D6 has the same problem. The sum clusters in the middle, so the middle options come up far more often. Option selection needs a flat distribution, so use a single die or apply the reroll rule strictly.

When the option count does not match a die's face count, the number picker is more accurate. It takes any range you specify and draws from it without bias.

Short runs look lopsided, and that is fine

Rolling three sixes in ten attempts does not mean the die is broken. Getting a specific face three or more times in ten rolls happens about twenty-two percent of the time.

Approaching the theoretical distribution takes hundreds of rolls. That is the law of large numbers, and the flip side of it is that small samples can deviate substantially.

Most of the moments where a game feels unlucky are exactly this short-run variance. Rolls are independent, so a bad streak gives no reason to expect a good one next.