Is Spin the Wheel Rigged? How Online Wheels Really Pick

Quick answer

An ordinary online spin-the-wheel is not rigged. It picks a winner with the browser's random number generator, every wedge is the same size, and each entry has the same chance on every spin. It often feels rigged because fair randomness produces far more streaks and gaps than people expect: on an eight-name wheel, the same name comes up twice in a row about one spin in eight. Rigged wheels do exist — this site has one, and says so in its name — and nothing on screen tells the two apart. The only honest test is counting results over many spins, and it takes more spins than most people think.

How a fair online wheel picks a winner

The wheels on this site work in two steps. First the code picks the winning entry: a random number between 0 and 1 is multiplied by the number of entries and rounded down, which gives every entry exactly the same share of the range. Only then does the wheel animate, turning a few full rotations and slowing down so that the pointer comes to rest on the wedge that was already chosen.

That order surprises people, because it feels as if the spin decides the result. It does not: the animation is a way of showing a decision, not of making one. A physical wheel is different: there, the result really does depend on how hard it was spun. On a screen, the number comes first and the motion second.

The random number comes from the browser. In JavaScript that is usually Math.random(), a pseudo-random generator: a formula that produces numbers with no pattern anyone can use, though it is not designed to resist a determined attacker. For choosing who goes first, picking a dinner or drawing a letter, that is entirely adequate. For anything that needs to stand up to scrutiny, such as a lottery, browsers also offer a cryptographic generator, and real prize draws should be run with a documented process rather than a web page anyway.

Why a fair wheel looks rigged

People have a strong sense of what random "should" look like — evenly spread, no repeats, everyone getting a turn — and real randomness is lumpier than that. Take a wheel with eight names:

• Back-to-back repeats are common. The chance that the next spin lands on the same name as the last one is 1 in 8, or 12.5%. Over ten spins, the chance of at least one back-to-back repeat somewhere is about 70%.
• Someone usually gets left out. The chance that one particular name never comes up in ten spins is about 26%. The chance that every one of the eight names appears at least once in ten spins is under 3%.
• Hot streaks happen. The chance that one particular name wins three or more of ten spins is about 12% — and with eight names in play, it is quite likely that somebody has a run like that.

So when a classroom wheel picks the same student twice in a lesson, or a friend's name never comes up all evening, that is what a fair wheel does. It is not evidence of anything on its own.

How a rigged wheel works

A rigged wheel uses exactly the same animation. The only change is in the first step: instead of choosing the winner at random, the code is told which entry to land on, then spins to it. Because the stopping point inside the wedge and the number of turns still vary, the motion looks just as natural as a fair spin.

The rigged wheel spinner on this site offers four ways to fix the result:

• Always lands on — the same entry wins every spin.
• Usually lands on — one entry wins a set share of spins, from 50% to 90%, and the rest are fair.
• Never lands on — the wheel skips any entries you tick and picks fairly among the others.
• Set the order — you queue the winners, and each spin takes the next one.

The controls can be hidden, and the wheel can go full screen, so the people watching see an ordinary wheel. That is the whole point of the tool, and also the reason it is labelled as rigged in plain words: it is for pranks, magic tricks, scripted demos and lessons on how easily "random" can be faked, not for anything people believe is fair.

How to tell whether a wheel is rigged

You cannot tell from one spin, or from three. What you can do is count. On a fair wheel of eight names, each name should win roughly one spin in eight over a long run, and a name that never wins becomes less and less likely the longer you watch:

| Spins watched | Chance a fair 8-name wheel never picks one particular name |

|---|---|

| 10 | about 26% |

| 20 | about 7% |

| 30 | about 1.8% |

| 40 | about 0.5% |

| 60 | about 0.03% |

Ten spins without a name is nothing unusual. Forty spins without it happens about once in two hundred tries, which is the point where asking a question is reasonable. The same logic works the other way: one name winning far more than its share over dozens of spins is worth noticing, while one name winning twice in a row is not.

Three practical checks go further than any statistics:

• Spin it yourself, on your own device. A wheel running on someone else's screen can be anything.
• Look at who controls the list. Duplicate entries are the oldest trick of all: a name typed twice has twice the chance, and on a crowded wheel nobody notices.
• For anything that matters, do not use a screen at all. Draw names from a hat in front of everyone, or use a process the people affected have agreed to in advance.

When rigging is fine, and when it is not

A fixed wheel is harmless when everyone will be told, or when nothing depends on the result: a prank that ends in a reveal, a magic trick, a demo that has to match a script, or a teacher showing a class that an interface can look random while being nothing of the sort. It stops being harmless the moment someone makes a decision, hands over money or enters a draw believing the result was down to luck. In many places, promotional draws are covered by consumer-protection or lottery rules, and "the wheel chose" is not a defence when you chose.

Try it both ways

Open the fair wheel spinner, add eight names and spin it twenty times, counting the results. Then do the same with the rigged wheel spinner set to "Never lands on" for one name. The two runs will look equally convincing on screen. Only the count will tell them apart — and even then, only if you spin long enough.