Weighted Odds, Explained
August 2, 2026
By default every name on a wheel gets an equal slice and an equal chance. Weighting changes that on purpose. It is a simple mechanism with one common misreading, and it is worth understanding exactly what a “weight” does before you rely on it in front of an audience.
A worked example
Say a wheel has three names: Alex with a weight of 1, Sam with a weight of 1, and Jordan with a weight of 3. The wheel does not split it three ways evenly. It adds the weights together (1 + 1 + 3 = 5) and gives each name a slice proportional to its share of that total:
- Alex: 1⁄5 of the wheel, a 20% chance
- Sam: 1⁄5 of the wheel, a 20% chance
- Jordan: 3⁄5 of the wheel, a 60% chance
Jordan is three times more likely to be picked than Alex or Sam, matching the weight exactly, not just “more likely” in some vague sense.
The mistake: a weight is not a percentage
This is the misreading, and it is worth stating flatly. Giving somebody a weight of 3 does not give them a 3% chance, or a 30% chance, or any fixed chance at all. A weight only has meaning relative to the total of every other weight on the wheel.
Which means the same weight produces completely different odds depending on what else is there. Jordan’s weight of 3 is a 60% chance against two names at 1 each. Put Jordan on a wheel with twenty other names at 1 each and that same weight of 3 is now 3 out of 23, or about 13%. Nothing about Jordan changed.
The practical consequence: adding names changes existing odds. If you promise somebody “triple chances” that promise survives new entrants, because it is a statement about the ratio. If you promise them “a 60% chance”, it does not survive the next person joining. Promise ratios, not percentages.
Think in tickets
The clearest mental model is a raffle box. A weight is a number of tickets. Total the tickets, draw one at random, and whoever holds it wins. Somebody with 3 tickets in a box of 5 wins three times out of five.
This is not just an analogy, it is what the code does. The draw sums every participant’s weight, picks a random integer in that range, then walks the list subtracting weights until it goes negative. The name it stops on holds the winning ticket. Thinking in tickets will never mislead you about the arithmetic.
Working backwards from the odds you want
Usually you know the target and need the weights. The method is to express the odds you want as a ratio and clear the fractions.
Want one name at 50% against three others sharing the rest? The favoured name needs as many tickets as the other three combined. Three names at 1 each is 3 tickets, so the favoured name gets 3, giving a total of 6, and 3⁄6 is exactly half.
Want roughly 25% for one name in a field of four? That is already equal weights, so leave it alone. Want 40/30/30 across three names? Use 4, 3 and 3, giving a total of 10, and the percentages fall out directly. When the target percentages sum to 100 and share a common factor, the weights are just those numbers.
Weights below 1, and zero
There is no way to give somebody a weight of zero to exclude them, and no fractional weights: anything below 1 is treated as 1. That is deliberate, because a zero-weight participant would be an invisible slice on the wheel, present in the list, impossible to win, and impossible for anyone watching to detect.
To take somebody out of a draw, take them off the wheel. Bench them for a round, which keeps their history and win count intact, or remove them. Either is visible; a silent zero would not be.
Weights versus listing a name twice
The free wheel has no weight field. Instead you list a name more than once, and two entries means two slices and twice the chance. Mathematically it is identical to a weight of 2, being the same number of tickets in the box.
One behavioural difference matters, and it catches people out. With elimination turned on, a win removes one slice, not the name. Somebody listed twice who wins still has their second slice and can be drawn again. That is usually exactly right, since it is what extra raffle tickets should do, but it surprises anyone who expected a single win to take them out entirely. If you want a name removed on its first win, list it once and weight it some other way, or take it off manually.
The slice size is the odds, not a decoration
The wheel is drawn to scale from the weights, so what you see is what you get: a slice that looks three times bigger really is three times more likely to win.
This does more work than it appears to. Every other way of weighting a draw, whether a spreadsheet column, duplicate rows in a list, or a tiering scheme described in a caption, makes the odds a claim the audience has to accept. On a wheel the odds are a picture, and anyone watching can check it against what you told them. It is the only method in the comparison of randomisers that displays its own probabilities.
A consequence worth planning for: weighting is visible whether or not you mention it. If you weight a public draw and say nothing, you have not concealed anything. You have just let people discover it themselves, which is worse than announcing it. Say it up front.
Keep the numbers legible
Weights of 1, 2 and 5 are numbers an audience can hold in their head. Weights of 7, 13 and 22 are arithmetic nobody in the room is going to do, and a tiering scheme with fourteen levels is one nobody can check, which forfeits the entire advantage of drawing the odds to scale in the first place.
If the tiers are getting complicated, that is usually a sign the giveaway rules are too complicated. Simplify the rules rather than encoding the complexity in weights.
Reasons to use it
- Extra raffle entries, for somebody who shared a post or referred a friend
- Handicaps, giving quieter students better odds in a classroom pick without excluding anyone
- Genuine preference in a decision, leaning towards an option without ruling out the rest. See the decision maker guide
- Tiered giveaway entries. See giveaway ideas
What it does not change
Weighting adjusts the odds up front, visibly, before the spin. It has nothing to do with how the winner is drawn. The pick still happens on the server using a cryptographic random source, exactly as on an unweighted wheel; only the number of tickets going into the draw is different. See how the live spin actually works for that part, or how random number generators actually work for what “cryptographic random source” is doing in that sentence.
Setting it up
On the free wheel, list a name as many times as it has tickets. On an account-backed wheel, set a weight per name. The how-to guide covers where.