Is a random draw fair? Unique results and repeats explained

The switch marked "unique results" changes the odds of every draw after the first. Here is what it does, in plain terms.

Every round of the generator picks animals from a pool. There is one setting that matters for fairness, and it is the switch labelled unique results. It is on by default. When it is on, no animal appears twice inside the same round. When it is off, each pick is independent and the same animal can come up twice in a row.

What changes when unique results is on

With unique results on, the draw works like dealing cards. The first animal is removed from the pool once it is picked, so the second pick cannot be the same one. With 7,119 species and a round of five, that difference is invisible. With a short custom list and a long round, it is the whole behaviour: asking for ten animals from a list of six returns six, not ten.

With unique results off, the draw works like rolling dice. Each pick is independent, so a repeat is possible. For a single pick — the common case for a class draw — the switch makes no difference at all, because there is nothing to repeat.

Why the first pick is equally likely

A fair draw means every entry in the pool has the same chance of being chosen. That is true here for the first pick of a round: the pool is not weighted, it is not sorted by popularity, and it is not ordered by anything you can exploit. A round of one is therefore as fair as this kind of tool can be.

Two things do change the odds, and both are choices you make before the draw:

The question a class always asks

"Why did we get the same group twice?" Two different things cause this, and they are worth separating.

The first is real repetition: the same animal came up in two separate rounds. This is normal and expected. With a pool of 7,119 and two rounds of one, the chance of an exact repeat is about one in seven thousand, which is small but not zero — and across a school year, with hundreds of draws, small probabilities happen.

The second is perceived repetition: two different animals from the same group, which feel like the same result. A cheetah and a serval are both cats; a newt and a salamander look alike to most people. Nothing has gone wrong. The pool is fine; the categories in our heads are coarser than the pool.

Making a classroom draw defensible

Fairness in a classroom is mostly about procedure, not mathematics. Four habits cover almost every objection:

  1. Draw in front of the class, once. Not before the lesson, not twice with the second result kept. The visible single draw is the thing that makes it credible.
  2. Say the pool out loud before the draw. "This is from all 7,119 species" and "this is from the twelve garden birds we listed" are different promises, and students accept both if they hear them first.
  3. Write the result down immediately. A result that is not recorded is a result that will be argued about, especially if it is unpopular.
  4. Decide in advance what a repeat means. If the same animal comes up twice in a term, is that allowed? Decide before it happens, because deciding afterwards always looks like a fix.

What the tool does not do

It does not record past draws, so it cannot avoid repeating an animal that came up last week. It does not offer a seed, so a draw cannot be reproduced exactly. It does not audit itself, so there is no log to show a sceptical student.

If you need any of those three things — no repeats across a term, a reproducible draw, a written record — keep the record yourself. A sheet of paper with the date and the drawn animal on it covers all three, and it takes less time than explaining why the computer is not cheating.

Related