Utilities
Dice Roller
Dice Roller generates something unpredictable: numbers, passwords, text or codes.
What this does
Dice Roller generates something unpredictable: numbers, passwords, text or codes.
Dice Roller produces random numbers, dice rolls, coin flips, and shuffled lists for games, sampling, giveaways, and simulations. Randomness here means each possible outcome is equally likely and independent of the last: a die has no memory of what it rolled before, and neither does a good generator.
The formula, explained plainly
Computers generate randomness with pseudorandom number generators (PRNGs): deterministic algorithms that produce sequences so statistically even that they pass as random for everyday purposes. Given the same starting seed they reproduce the same sequence, which is a feature, not a bug: it lets you replay a simulation exactly.
Uniform distribution is the core promise: over many draws, each outcome appears about equally often. In small samples randomness looks lumpy, with streaks and clusters that feel meaningful but are not. This lumpiness is the single biggest source of 'this generator is rigged' complaints, and it is completely normal.
Dice, coins, and draws are probability made tangible. Two dice have 36 equally likely outcomes, and 7 is the most common total because six of those outcomes make 7. Sampling tools extend the idea: drawing with replacement (the item goes back) keeps odds constant, while drawing without replacement (names from a hat) changes the odds with every draw.
Shuffling and random ordering use the Fisher-Yates method conceptually: walk through the list swapping each item with a randomly chosen later item, which gives every ordering an equal chance. Dice Roller applies the same idea so raffles, teams, and playlists come out fair.
How to use it
- Enter dice.
- Enter sides.
- Read the instant result and the breakdown below it.
- Adjust any input to compare scenarios.
Worked example
With dice = 2, sides = 6, the result is Total 10 Rolls is 4, 6.. Outputs change every run, by design.
Common mistakes
- Using a general-purpose random generator for passwords, tokens, or gambling, where cryptographically secure randomness is required.
- Reusing the same seed and wondering why every 'random' run gives identical results: that is what seeds are for.
- Introducing modulo bias by taking rand() % n when n does not divide the generator's range evenly (matters for serious simulations).
- Drawing samples too small to be representative, then treating the result as a reliable estimate.
- Confusing permutations (order matters) with combinations (order does not) and computing the wrong count of outcomes.
- Assuming a shuffled playlist with no repeats is 'less random' than one that sometimes repeats songs: true randomness repeats.
- Testing fairness on a handful of draws instead of thousands, where the statistics actually settle.
- Forgetting that 'random' selection from a sorted or filtered list first requires the list itself to be complete and unbiased.
Limitations
- The tools model fair, independent draws; loaded dice, biased coins, and real-world selection biases are outside the model.
- Very large shuffles and samples are limited by browser performance and memory, not by the math.
- Probability outputs describe the process, not a guarantee about your next specific draw.
- Browser generators are pseudorandom, not cryptographically secure: fine for games and sampling, not for secrets or real-money gambling.
- Without a fixed seed, sequences are not reproducible; with a fixed seed, they are not surprising. Pick the property you need.
Expected accuracy
Generators use the browser's built-in pseudorandom source (xorshift-based in modern engines), which is statistically uniform for non-cryptographic use. Dice, card, and sampling tools draw from exact combinatorial counts, so stated probabilities (like 6/36 for rolling 7) are mathematically exact.
Privacy
Everything you type stays on your device. The calculation runs in your browser with JavaScript; no input is sent to a server, stored in an account, or shared with anyone.
Sources and standards
- Standard PRNG behavior as implemented in JavaScript engines; dice and sampling probabilities from classical combinatorial definitions (permutations, combinations, uniform distribution).
Bottom line
Dice Roller gives you fair, uniform randomness for games, draws, and everyday simulation, with optional seeds when you need a replayable sequence. Expect streaks and clusters in short runs: that is what real randomness looks like. For passwords, tokens, or anything adversarial, switch to a cryptographically secure generator instead.
Key insight
Humans are terrible at randomness: we see patterns in noise and distrust true random sequences because they contain streaks. If your random draw looks streaky, that is evidence it is working.
Frequently asked questions
How do I generate a random password?
Use a cryptographically secure generator with enough length: 16+ random characters from a large set. Do not use Dice Roller for this: its pseudorandom output is fine for games but not designed to resist someone trying to guess your secrets.
What are the odds of winning the lottery?
For a 6-from-49 lottery, 1 in about 13.98 million per ticket. Buying 100 tickets makes it 100 in 13.98 million, still about 1 in 140,000. Dice Roller computes exact odds for any draw format so the scale of the long shot is visible.
Why does randomness feel 'clumpy'?
Because even spacing is the unnatural pattern: if you asked people to fake 100 coin flips, they alternate too regularly and avoid long runs. Real randomness clusters. When a generator's output looks too smooth to be random, that is actually the suspicious case.
Is this truly random?
It is pseudorandom: a deterministic algorithm with a statistically uniform output. For games, raffles, sampling, and simulations it is indistinguishable from true randomness. For passwords, encryption, or gambling, use a cryptographically secure generator instead.
What is a seed and why would I set one?
A seed is the starting value of the generator. The same seed reproduces the exact same sequence, which makes simulations replayable and bugs reproducible. Leave it unset (or random) when you want a fresh, unpredictable sequence each time.
Why do I keep seeing streaks?
Because streaks are normal in random sequences. In 100 coin flips, a run of 6 heads in a row is actually expected. Our brains are wired to see streaks as meaningful, but in independent random draws they carry zero predictive information.
What are the odds of rolling 7 on two dice?
6 in 36, or about 16.7%: the most likely total. Six combinations make 7 (1+6, 2+5, 3+4, 4+3, 5+2, 6+1) out of 36 equally likely ordered pairs. Totals of 2 and 12 have only one combination each, about 2.8%.
Is a coin flip really 50/50?
Mathematically yes for a fair coin. Physically, real coins have a microscopic bias toward landing on the face that started up (about 51/49 in large studies). For every practical purpose, treat it as 50/50 and let Dice Roller do the flipping.
How do I pick a fair random winner?
Put every entrant in the list exactly once, shuffle or draw with Dice Roller, and take the top result. The fairness lives in the complete, unbiased list plus a uniform draw: no generator can fix a list with duplicates or missing names.
What is sampling with vs without replacement?
With replacement, each draw goes back, so odds never change (like rolling a die repeatedly). Without replacement, each draw removes the item, so odds shift every time (like dealing cards). Raffles are without replacement; dice are with.
What is the difference between permutations and combinations?
Permutations count orderings (ABC differs from CBA); combinations count groups (ABC is the same as CBA). There are 6 permutations but 1 combination of 3 items. Use permutations for passwords and race finishes, combinations for lottery draws and committees.
Can random numbers be predicted?
Pseudorandom sequences can be predicted if someone learns the algorithm state, which is why they are unsafe for security. Without that inside knowledge, the next value is for all practical purposes unpredictable. Cryptographic generators are designed to resist even state-recovery attacks.
How many samples do I need for a fair estimate?
It depends on the precision you want: estimates tighten roughly with the square root of the sample size, so quadrupling samples halves the error. A few hundred draws give a decent picture of simple proportions; rare events need far more.