One die: every face is equally likely
A fair six-sided die has six faces, and each one comes up with probability 1/6, about 16.7%. A fair twenty-sided die gives each face 1/20, or 5%. That flat spread is the starting point for almost every dice calculation: count the faces that help you, divide by the total number of faces.
Needing a 5 or 6 on a d6 means two helpful faces out of six, so a one-in-three chance. Needing a 15 or higher on a d20 means six helpful faces (15 to 20) out of twenty, which is 30%. Each step on a d20 is worth 5 percentage points, a handy rule when a modifier of +1 or -1 changes what you need.
Rolls are independent
Dice have no memory. If a d6 has shown no sixes in ten rolls, the next roll still has a 1/6 chance of a six. The feeling that a result is “due” is one of the most common misreadings of random events, and it matters because it nudges people toward bad choices, such as holding on to a risky plan because the dice “owe” them. What the dice can tell you is how likely something is before the roll, not what the previous rolls imply about the next one.
Adding dice changes the shape
Roll two six-sided dice and add them, and the results stop being flat. There are 6 x 6 = 36 equally likely ordered outcomes, but the totals share them unevenly:
| Total | Ways to roll it | Probability |
|---|---|---|
| 2 or 12 | 1 each | 1/36 (about 2.8%) |
| 3 or 11 | 2 each | 2/36 (about 5.6%) |
| 4 or 10 | 3 each | 3/36 (about 8.3%) |
| 5 or 9 | 4 each | 4/36 (about 11.1%) |
| 6 or 8 | 5 each | 5/36 (about 13.9%) |
| 7 | 6 | 6/36 (about 16.7%) |
The ways add up to 1+2+3+4+5+6+5+4+3+2+1 = 36, as they should. A seven is the most common total because there are six combinations that make it (1+6, 2+5, 3+4 and their reverses), while a twelve needs two sixes.
This is why games that ask for the sum of two dice feel very different from games that use one d12. A d12 gives every number from 1 to 12 the same 1/12 chance. Two d6 cluster results around the middle, so extreme totals are rare and the middle is reliable. If a game gives you a choice between those two options, you are really choosing between volatility and consistency.
More dice, a tighter bell curve
Add a third die and the effect grows. Three six-sided dice have 6 x 6 x 6 = 216 outcomes. A total of 10 or 11 can each be made 27 ways, so 27/216 or 12.5% each, while 3 and 18 can each be made only one way, 1/216, under half a percent. Games that use several dice to measure a skill or a result are, in effect, deciding how much surprise to allow. Understanding those outcome distributions helps you see which results are routine and which are genuinely remarkable.
“At least one” and the complement trick
Many dice questions are easier if you calculate the chance that something does not happen and subtract it from 1. The chance of at least one six in four rolls of a d6 is 1 minus the chance of no sixes at all:
1 - (5/6)^4 = 1 - 625/1296 = 671/1296, about 51.8%.
So four rolls give a little better than even odds of seeing at least one six. Notice that the answer is not 4 x 1/6 = 66.7%. Adding the chances together double-counts the cases where several sixes appear, which is a very common mistake.
The same idea explains rolling with advantage in games that allow it: roll two d20 and keep the higher. If you need a 15 or more, the chance that a single die fails is 14/20 = 0.7, so both failing is 0.7 x 0.7 = 0.49, and success is 51% instead of 30%. If you need an 11 or more, a single roll succeeds 50% of the time and advantage lifts that to 1 - 0.25 = 75%. Advantage helps most when the base chance is middling, and by a smaller amount when the target is already very easy or very hard.
What this means for strategy
The distribution of results does not tell you how to win, but it does shape sensible decisions.
- Compare chances, not hopes. If one action succeeds on a 9 or higher with two d6 (10/36, about 28%) and another on a 5 or higher on one d6 (2/6, about 33%), the second is somewhat likelier, even though the first “feels” bigger.
- Count the cost of failing. A 30% chance of a great result might be worth taking if failure costs little, and a 75% chance might be poor if failure ends the game. This kind of risk assessment is as important as the probability itself.
- Prefer reliability when you are ahead and variance when you are behind. Games that let you choose between steady and swingy dice give you a way to do that, though it never guarantees an outcome.
- Do not expect the long run to arrive quickly. Averages show up over many rolls, but a handful of rolls can stray a long way from them. A short streak says very little about whether dice are fair.
Some people hope that knowing the numbers will let them go as far as beating the odds in a game of pure chance. It does not. In a game where the dice decide the result and the odds are fixed, the numbers help you understand and compare risks, not remove them. In games where you choose what to do with the dice, probability helps you pick between options, but outcomes remain uncertain.
Checking whether dice are fair
A die that is slightly off would show up only across a large number of rolls. A few dozen rolls cannot tell you much, because uneven-looking results are normal in small samples. If you want to test a die, roll it many times, record the results and compare each face’s count to the expected share. Even then, small differences are usually just chance.
Where to go from here
Most dice games combine the ideas above in different ways: a single flat roll, a summed pair, a pool of dice, or a reroll option. Once you can count the combinations for each, the rest is comparing what each choice is likely to give you and what it risks. Dice results are the clearest place to practise this, because every outcome can be listed and counted by hand.