Scratch cards and instants

Why the Odds Printed on the Back Are Not the Odds You Face

The printed figure is accurate and misleading at the same time — because it measures how many tickets win, not how much they win.

Every scratch card carries a line like Overall odds of winning: 1 in 3.63. It is not a lie, it is not out of date, and it is not what you want to know.

What the printed figure actually measures

It measures the proportion of tickets in the complete print run that win any prize at all, computed at design time:

Overall odds = tickets printed ÷ total winning tickets

In the worked game from how scratch card odds work — 4,000,000 tickets, 1,101,022 winners — that is 1 in 3.63.

Two things about this number matter, and both are usually missed.

Problem 1: it counts winners, not value

The overall odds treat a $5 prize and a $500,000 prize identically. Both are "a win".

In that game, 920,000 of the 1,101,022 winning tickets — 84% of all winners — pay exactly $5, which is the ticket price. Winning that prize returns your stake and nothing more.

So "1 in 3.63 tickets wins" is compatible with "most winning tickets do not put you ahead". A figure that counts a refund as a win is measuring something other than what a player means by winning.

Problem 2: it barely changes even as the game degrades

This is the subtle one and it is what makes the figure genuinely misleading late in a game's life.

Take that game half-sold, with both $500,000 top prizes claimed and everything else claimed proportionally:

  • Prize money remaining: $5,800,000 in 2,000,000 tickets → EV falls from $3.40 to $2.90, a return of 58% rather than 68%.
  • Winners remaining: about 550,509 in 2,000,000 tickets → overall odds 1 in 3.63.

The odds figure has not moved. The expected value has dropped ten points. Because the top prizes are a tiny fraction of the count of winners but a real fraction of the value, removing them guts the EV while leaving the printed ratio essentially untouched.

The printed odds therefore remain accurate for the life of the game and stop describing what you are buying almost immediately.

What to look at instead

Prizes remaining data. Most US state lotteries publish, per game, how many of each prize were printed and how many are left. That is the information the printed odds cannot carry, because it did not exist at print time — see how to check prizes remaining.

Expected value, computed now. EV = Σ (prize × prizes remaining) ÷ tickets remaining. Method and caveats in scratch card expected value; the scratch EV calculator does it for you.

The game's payout percentage. Operators publish this by price point — Minnesota's range from 63% at $1 to 74% at $50. It tells you far more than the overall odds do (which price point has the best return).

Is it deliberate?

Not in the sense of dishonesty. Overall odds are a printed artefact: the card is manufactured before a single ticket is sold, so nothing dynamic can be on it, and regulators generally require overall odds to be disclosed. The figure is the most informative thing that can be fixed at print time.

The problem is not the disclosure, it is the framing that grows around it. "1 in 3.63 wins!" is marketing that is technically true and carries an implication — that a win is a gain — which the prize structure contradicts for 84% of winners.

Compare how the same asymmetry works in draw games, where operators quote the odds of any prize rather than the jackpot: 1 in 24.9 sounds much better than 1 in 292,201,338, and the small print is the same small print.

The one-line version

The printed odds tell you how often a ticket wins something. They tell you nothing about how much, and nothing about whether the good prizes are still in the box.

Try it yourself

Keep reading

Sources

Last verified: 2026-08-29