How randomness works here

Every generator on this site draws from your browser’s cryptographically strong random source, exposed as crypto.getRandomValues. Nothing is generated on a server, and no result is derived from the time of day, your activity or a fixed seed.

A random machine integer has to be mapped into the range you asked for. Doing that with plain remainder arithmetic makes the first few values in the range come up slightly more often — the classic modulo bias.

Instead, raw values that fall outside the largest evenly divisible interval are discarded and a new value is drawn. The remainder is only taken once the value is known to be in a region that divides the range evenly, so every value in the range is equally likely. Large ranges are handled with big-integer arithmetic so the same guarantee holds at the edges of the supported range.

Lists are shuffled with a Fisher-Yates shuffle whose random index at each step comes from the same unbiased integer generator. Teams shuffle once and deal round-robin, so group sizes differ by at most one. Pairs shuffle once and take neighbours.

Without replacement means a record can be drawn at most once — this is what most people mean by taking a sample. With replacement means each draw is independent, so the same record can come up more than once. Small samples from big lists use a partial Fisher-Yates pass rather than shuffling everything.

Random dates are chosen from the set of whole calendar days in your range, not by picking a point between two timestamps. That matters because daylight-saving changes make some local days 23 or 25 hours long, which would otherwise give those days slightly different odds. Day filters build the eligible set first and sample from it, rather than drawing repeatedly until a matching day appears.

Randomness cannot fix a biased source list. If a name appears twice, it has two entries and roughly twice the chance of being picked. Duplicates are kept because they are often intentional, and each list tool tells you how many values repeat.

These tools are not certified, audited or “provably fair”, and no software can promise identical behaviour in every possible browser or runtime. What we can say is concrete: results use your browser’s cryptographically strong random source, range mapping is unbiased, and unweighted draws give each eligible item an equal chance.

These tools are also not suitable for generating passwords, encryption keys, authentication secrets, recovery phrases or cryptocurrency keys. Those needs have different requirements.

These tools are intended for everyday selection, games, teaching, sampling and general utility. Check applicable rules before using random draws for regulated lotteries, gambling, legal selection processes, scientific protocols or other regulated, high-stakes decisions.