Dice Roller & Coin Flip
Roll any dice from d4 to d100, or flip a coin — no physical dice needed.
Results come from your browser’s cryptographic random number generator, with rejection sampling so every face is equally likely.
Roll any standard polyhedral die from a d4 to a d100, or flip a coin, with running totals, averages and a heads/tails tally kept as you go — no physical dice needed, and none to lose under the sofa.
Every roll uses your browser's cryptographic random number generator rather than a simple pseudorandom function, so the distribution is genuinely uniform and no outcome is more likely than any other.
The probabilities behind the dice
A single die is uniform — every face is equally likely. Add dice together and that stops being true, because there are many more ways to roll a middling total than an extreme one.
| Total on 2d6 | Ways to roll it | Probability |
|---|---|---|
| 2 or 12 | 1 | 2.8% |
| 3 or 11 | 2 | 5.6% |
| 4 or 10 | 3 | 8.3% |
| 5 or 9 | 4 | 11.1% |
| 6 or 8 | 5 | 13.9% |
| 7 | 6 | 16.7% |
There are six ways to make 7 and only one to make 12, which is why 7 is the pivot of every dice game ever designed around 2d6.
Notation you'll meet in tabletop games
- 3d6 means roll three six-sided dice and add them. The average is 10.5.
- 1d20+5 means roll one twenty-sided die and add five — the core of most role-playing skill checks.
- 2d6+3 has a range of 5 to 15, with 10 the most likely single result.
- 'Advantage' means rolling two d20s and keeping the higher, which raises the average from 10.5 to about 13.8.
- 'Disadvantage' is the mirror image: roll two and keep the lower, dropping the average to about 7.2.
- d100 (or percentile) is used where a flat 1-in-100 chance is wanted, with no bell curve at all.
Why streaks feel impossible but aren't
Dice have no memory. After four sixes in a row, the chance of a fifth is still exactly one in six — the die does not know what it did before. The strong intuition that a result is 'due' is the gambler's fallacy, and it is one of the most reliably wrong instincts people have.
Genuinely random sequences also contain far more clustering than people expect. In 100 coin flips, a run of six or more identical results appears more often than not. When people are asked to fake a random sequence they produce something too evenly alternating — which is exactly how researchers detect fabricated data.
Frequently asked questions
- Are the rolls genuinely random?
- They use your browser's crypto.getRandomValues, the same cryptographic-quality source used for security purposes, rather than a basic pseudorandom function. Every face has an equal chance.
- Can I roll multiple dice at once with a modifier?
- Yes — set the number of dice and add a modifier, and the tool shows both the individual results and the combined total.
- Does it keep a history of my rolls?
- Yes, a running list with totals and averages is kept for the session. It's stored only in your browser and cleared on reload.
- I rolled five sixes in a row — is it broken?
- No. That sequence has a probability of about 1 in 7,776, which sounds rare but happens regularly across thousands of rolls. Dice have no memory, so a sixth six remains a 1-in-6 chance.
- Is a virtual die as fair as a physical one?
- Arguably fairer. Real dice have manufacturing imperfections, rounded edges and weight asymmetries — cheap plastic dice measurably favour some faces. A cryptographic generator has no such bias.
- What's the average roll on 3d6?
- 10.5. Each d6 averages 3.5, so three of them average 10.5, with results clustering around that figure far more often than at the extremes of 3 or 18.