Scratch games

Scratch-off tickets are sold on a number printed on the front. What that number never tells you is whether the prize is still there. A game whose top prizes have all been claimed sits on the same rack, at the same price, advertising the same jackpot.

State lotteries publish enough data to settle it — how many of each prize were printed and how many have been claimed — so we do the arithmetic for every game on sale: how much prize money is genuinely left, how many tickets remain, and therefore what a dollar spent on that game buys today rather than at launch.

The one finding that holds everywhere

Across every state we've analysed, the same pattern appears: the more a ticket costs, the more of each dollar comes back as prizes. We can now test that in eighteen state lotteries that publish enough data to price their games. It holds in all eighteen — there is no counter-example:

• Texas 55% → 78%  ·  California 61% → 83%  ·  Georgia 62% → 80%  ·  North Carolina 57% → 83%  ·  Virginia 58% → 81%
• Missouri 61% → 79%  ·  South Carolina 58% → 77%  ·  Connecticut 62% → 80%  ·  Massachusetts 67% → 81%  ·  Oklahoma 56% → 83%
• New Mexico 53% → 69%  ·  New Jersey 59% → 74%  ·  Maryland 55% → 79%  ·  Arizona 58% → 78%  ·  Arkansas 60% → 78%
• Louisiana 57% → 73%  ·  Illinois 65% → 82%  ·  Idaho 69% → 91%

Each pair is the average value returned per dollar on that state's cheapest tickets against its most expensive. Eighteen lotteries, none of which sets its prize structures with reference to the others, all making the same design decision: the cheapest ticket on the rack is reliably the worst value in the shop.

We wrote up the full analysis here — all 1,275 games, why cheap tickets are built this way, and why it is still not a reason to buy the expensive ones.

New York is the exception, and deliberately so: it publishes prize data but not ticket prices, so rather than guess we rank its games by a price-independent index of how much prize money is still unclaimed. The method differs; the question it answers is the same.

It is worth being precise about what that does and doesn't mean. A more expensive ticket is a smaller loss per dollar staked — it is not a good bet, and it puts far more money at risk per ticket. Every game on every page here returns less than it costs across all players. These tables show which games are currently least bad, never which are good.

How the analysis works

For each game we take the published prize table — every tier's original count and how many remain unclaimed — and the size of the print run, either stated outright or derived from prize counts times their published odds and cross-checked across tiers. Unclaimed prize value divided by the cost of the remaining tickets gives value per dollar right now; the same sum across the whole print run gives what the game paid at launch. The gap between the two is what the rack never shows you.

Two limitations apply everywhere. Percent sold is inferred from claims where a state doesn't publish it, which understates sales slightly and makes our figures conservative. And "remaining" means unclaimed, not unsold — a top prize may already sit in a bought ticket, which is why games more than 90% sold are flagged and hidden by default. Full detail sits on each state page and in our methodology.

For draw games rather than scratch, see US National Drawings, US State Drawings, or the most winnable game in every state.