Is a weighted draw still a fair draw?

Yes, if the weights follow a rule that was published before entries closed and that anyone can apply to themselves. A draw is unfair when the weights encode a preference nobody was told about, not when they are unequal. Two tickets for a referral is fair; two tickets because you like someone is not, however random the draw afterwards.

Does a weight of two mean twice the chance of winning?

Only when you are drawing a single winner. As soon as you draw several winners without replacement, everyone’s probability rises towards certainty and the gap between weighted and unweighted entrants compresses. In a three-entrant pool where one holds two tickets, that entrant is twice as likely to win a single prize but only about 1.4 times as likely to take one of two prizes.

How much weight is too much?

Watch the share, not the multiplier. If one entrant holds more than a fifth of the tickets, the draw is mostly about them and other entrants can feel it even without doing the arithmetic. A ceiling on tickets per person, published with the rules, keeps a referral mechanic from turning a draw into a formality.

Is adding a name twice the same as giving it a weight of two?

For a single winner, yes. For several winners drawn without replacement, no — a duplicated name can be drawn twice and take two prizes, while a weighted entry is removed from the pool once it wins. Decide which of those you meant before the draw rather than discovering it when one person wins two prizes.

Weighting a draw without rigging it

Extra tickets for a referral, a second entry for last year's near-miss, more weight for the people who did the most work. Weighting is usually an attempt to be more fair, not less — and it is also the point at which a draw becomes hard to defend.

The line is disclosure, not equality

An unequal draw is not an unfair one. A raffle where some people bought five tickets and others bought one is unequal by design, and nobody objects, because the rule was obvious and available to everyone. What makes a weighted draw unfair is weights that encode a preference the entrants were never told about.

So the test is not are the chances equal. It is:

Was the rule published before entries closed?
Can an entrant apply it to themselves and work out roughly what they hold?
Was it fixed before you saw who entered?

That third one is where good intentions go wrong. Adjusting weights after the entry list exists — even to correct something genuinely unbalanced — means the weights are now a response to the entrants, and there is no way to show otherwise. If the rule needs changing, the honest move is to say so publicly and reopen entries, not to quietly revise the numbers.

Say it in one sentence, or do not do it

A weighting rule that cannot be stated in a sentence is one your entrants will not verify, and a rule nobody verifies is indistinguishable from no rule at all.

“One ticket for entering, one more for each friend who enters, up to five” is a sentence. “Weighted by engagement across the campaign period with adjustments for account age” is not — it is an assurance that a calculation happened somewhere, and trusting it means trusting you rather than checking anything. Write the sentence first and choose the numbers second; if it comes out clumsy, the scheme is too complicated for a draw.

Two tickets stops meaning double, and it happens fast

This is the part almost everyone gets wrong, and it is worth doing the arithmetic once because it is not intuitive.

Take three entrants. A holds two tickets; B and C hold one each. Four tickets in the pool.

Draw one winner. A wins two times in four — 50%. B and C get 25% each. A is exactly twice as likely to win, which is what everyone expects a double weight to mean.

Now draw two winners from the same pool, removing each winner as they come out. A now takes a prize about 83% of the time, and B and C about 58% each. A is still ahead — but the advantage has fallen from twice as likely to roughly 1.4 times as likely.

Nothing was rigged and no rule changed. The compression is arithmetic: as the number of prizes rises towards the number of entrants, everybody's chance of taking one rises towards certain, and there is less and less room for anyone to be ahead. Draw three winners from three entrants and the weights stop mattering entirely — everyone wins.

The practical consequence is that weighting does most of its work when prizes are scarce relative to entrants. If you are giving away twenty prizes to sixty people, a double weight is close to decorative, and an entrant who does the maths may reasonably feel the referral scheme promised more than it delivered. Either accept that, or say plainly that the extra tickets matter most for the single big prize.

Watch the share, not the multiplier

“Up to ten tickets” sounds modest. In a draw with fifteen entrants it means one person might hold nearly half the pool.

Check the largest share any single entrant holds before you run anything. Past roughly a fifth of the tickets a draw reads as a formality to everyone else, and that cost lands whether or not the weighted entrant wins. A published ceiling on tickets per person is the simplest fix, and it also caps anyone gaming a referral mechanic.

Duplicates and weights are not the same thing

Adding a name to the list twice and giving that name a weight of two are identical for a single-winner draw, and different as soon as there are several prizes.

A duplicated name can come out twice and take two prizes. A weighted entry is normally removed from the pool once it wins, so it takes one. Neither is wrong; they answer different questions, and the failure mode is discovering which one you chose when a single person wins two of your three prizes and you have to decide, in public and after the fact, whether that counts.

When the answer is not to weight at all

If you are reaching for weights because a particular person deserves to win, the honest thing is to choose them and say you chose them. Awarding a prize on merit is entirely respectable. Running it through a draw that was quietly tilted towards the outcome you wanted is the version that damages you if it ever comes out — and it usually comes out, because the tilt has to be large enough to work.

A draw is the right instrument when every entrant has an equal claim and you need a way to decide that nobody has to justify. Weighting extends that to cases where claims differ by a stated, checkable amount. Beyond that, you are not running a draw, and dressing a decision up as one costs more trust than making it openly ever would.

Publishing the weights, and weighting against someone

Should I publish everyone's ticket count?

Publish the rule and the ceiling; publish individual counts only if entrants knew that would happen. Anyone should be able to work out their own total from the rule. Nobody needs a public table of who referred the most friends, and producing one turns a giveaway into a leaderboard with a different set of arguments attached.

Can I weight against someone — a previous winner, say?

Yes, and it is often the fairest option in a recurring draw. Excluding previous winners outright is cleaner still, and easier to state. Whichever you pick, it belongs in the published rules from the first round rather than appearing in round three when the same person keeps winning.

Does weighting affect how random the draw is?

No. Weighting changes the shape of the distribution you are drawing from, not the quality of the randomness doing the drawing. A weighted draw from a good source is exactly as unpredictable as an unweighted one; it is simply unpredictable over a different set of odds. How the underlying numbers are produced is a separate question, and the random number generator sets out its own source and the bias correction it applies.

Weighting is one decision inside a larger procedure — the rest of it, from publishing rules to announcing reserves, is in how to run a prize draw people will accept as fair. To run a weighted draw: random name picker for names with tickets, or the random list generator for inline weights.