Why the same name keeps winning (and why that's normal)

You spin the classroom wheel, it lands on Jake. Two spins later, Jake again. Thirty faces turn towards you and someone says the thing they always say: it's rigged.

It isn't, and the reason is worth having ready, not just to defend the wheel, but because the misunderstanding underneath it is one of the most common in all of probability, and it takes about ninety seconds to clear up.

Random doesn't mean evenly spread

When people picture 'random', they picture something spread out: everyone taking a turn, no clumps, no repeats. That picture is of a shuffle, a deck dealt out, not of a random draw. A genuine random draw looks lumpier than almost anyone expects, and the lumps are the signature of the real thing.

Here's the key property: each spin is independent. The wheel keeps no record of who won last time. It cannot, because nothing about the previous result is stored anywhere; the pick is a fresh draw from the browser's cryptographic random source every single spin. So if you have 10 names, the chance the next spin repeats the current winner is 1 in 10. Not lower because they just won. Exactly 1 in 10, every time.

The numbers, so you can say them out loud

Take a 10-name wheel and 10 spins, a plausible lesson. What's the chance that every spin produces a different name, the tidy outcome people imagine is 'properly random'?

About 0.04%. Roughly one class in three thousand. Which means the chance that some name comes up at least twice is about 99.96%, which is practically a certainty. A run of ten spins with no repeats at all wouldn't be evidence that the wheel was fair; it would be the single most suspicious thing it could do.

Back-to-back repeats, the ones that actually cause the shouting, are common too. Across those same 10 spins, the chance of at least one name being picked twice in a row is about 61%. More often than not.

Why a no-repeat picker would be worse

The obvious fix is to make the picker refuse to repeat. It's worth seeing why that breaks the thing you wanted.

If Jake can't win twice in a row, then the moment Jake wins, his chance on the next spin becomes zero and everyone else's rises from 1-in-10 to 1-in-9. The picks now depend on history, which means they're partly predictable: you can rule out one name with certainty before the wheel even moves. You've traded a fair draw for a comfortable-looking one.

It also fails the test that matters in a classroom: it doesn't stop the complaint. Jake can still be picked on the third spin, and the fifth, and the ninth. Blocking only the immediate repeat removes the most visible clump while leaving the underlying clustering exactly as it was.

When you don't want randomness at all

Sometimes 'random' isn't actually what you need. If your goal is that every student answers exactly once, or every team presents once, you don't want an unpredictable draw; you want a shuffled order. Those are different tools and it's worth being deliberate about which one you're reaching for.

Turn on 'Remove winner from wheel after spin' and the wheel becomes a shuffle: each name comes off as it's drawn, so you work through the entire list with no repeats and nobody missed. Order is still unpredictable (nobody can tell when their turn is coming) but coverage is guaranteed. For cold-calling, presentation order, or working around a class, this is almost always the setting you want.

How long until everyone gets a turn?

If you leave repeats on and just keep spinning, waiting for everyone to be picked at least once takes much longer than people expect. For 10 names it's about 29 spins on average; for 30 names, about 120.

The reason is the tail. Early on, nearly every spin finds someone new. By the time only two names are left unpicked, most spins land on the 28 people already covered, and you're waiting a long time for the stragglers. If full coverage is the goal, don't wait for chance to deliver it; remove each winner and get there in exactly N spins.

Check it yourself

None of this needs to be taken on faith, and 'trust me, it's random' is a weak thing to say to a sceptical room. The probability simulator runs thousands of spins of your actual wheel and shows the distribution: the visible answer to 'does this favour anyone', and a genuinely good five minutes with a maths class.

What you'll see is the counterintuitive shape of it: over thousands of spins the counts flatten out towards equal, while any short stretch you zoom into still looks clumpy and unfair. Both are true at once. That's what randomness is.