Random Number Generator

A pseudorandom number generator (PRNG) is an algorithm that produces sequences of numbers which look statistically random but are actually computed deterministically from an internal seed state, rather than sampled from unpredictable physical noise. Every result here comes from JavaScript's Math.random(), a PRNG. This is more than sufficient for dice rolls, coin flips, raffles, and shuffling a playlist.

It is not appropriate for cryptography, security tokens, or real-money gambling, where a cryptographically secure PRNG (crypto.getRandomValues) or a hardware-based true random number generator sourced from atmospheric noise or radioactive decay (such as random.org) is required instead. The practical difference: a PRNG's output could theoretically be predicted if an attacker knew the internal seed state, which never matters for a game night but matters a great deal for generating a password or encryption key.

The List Randomizer uses the Fisher-Yates (Knuth) shuffle: it walks the list from the last item to the first, and at each step swaps the current item with a randomly chosen item from the remaining unshuffled portion, including itself. This runs in linear time and, critically, gives every possible ordering of the list an equal probability of occurring.

This matters because a naive approach, such as assigning each item a random number and sorting by it, is not perfectly uniform and can subtly favor certain orderings depending on the sort algorithm's stability. Fisher-Yates avoids that bias entirely, which is why it is the standard algorithm used in shuffling implementations across most programming languages.

Drawing one winner on screen for a class or a live stream? The spin the wheel random name picker shows the draw as a spinning wheel and can remove each winner.

For a single fair die with n sides, every face has an equal 1/n chance, and the expected (average) value across many rolls is (n+1)/2. A D6 averages 3.5, a D20 averages 10.5, and a D100 averages 50.5. Rolling multiple dice at once, as the tool's dice count option allows, changes the shape of the outcome: the sum of two D6 rolls is not evenly distributed like a single die but forms a triangular distribution peaking at 7, since there are more combinations that add up to 7 (1+6, 2+5, 3+4, and their reverses) than any other total.

This is useful for tabletop game design and probability practice: rolling several dice and summing the results narrows the range of likely outcomes toward the average, which is exactly why games use multiple smaller dice (2d6, 3d6) instead of one large die when they want more predictable, bell-shaped results rather than a flat, equally-likely spread.

Each coin flip or dice roll is an independent event, the generator has no memory of previous results. If a coin lands heads five times in a row, the probability of heads on the next flip is still exactly 50%, not lower. This misconception, expecting a run to "even out," is known as the gambler's fallacy. The same applies to the Random Number tab: if you have generated 1, 2, and 3 from a 1-10 range, 4 is not more or less likely on the next draw than it was on the first.

Long streaks are also less unusual than intuition suggests. With a fair coin, the chance of at least one run of 5 heads-or-tails in a row somewhere across 100 flips is well over 90%, streaks are an expected feature of randomness, not a sign that something is wrong with the generator.