Scratch cards and instants
The single most useful piece of public data in the lottery world, and how to turn it into an expected value rather than a vibe.
Instant games are the one lottery product where the operator publishes enough information to work out what you are actually buying. Most players never look.
Most US state lotteries maintain a prizes remaining page for every active scratch game, typically showing:
Some also publish the total number of tickets printed and the game's overall odds. That is enough to compute a genuine expected value.
Practice elsewhere varies. Some non-US operators publish top-prize availability only; some publish nothing beyond the launch odds. Where the data is not published, the honest answer is that you cannot evaluate the game — and you should treat any site claiming to have done so for that jurisdiction with suspicion.
Expected value of the tickets currently out there is:
EV = Σ (prize amount × prizes remaining) ÷ tickets remaining
The numerator is straightforward from the published table. The denominator is the hard part, because operators rarely publish tickets remaining directly.
The standard estimate uses the claim rate as a proxy. If a game printed 4,000,000 tickets with 1,101,022 winners and 40% of prizes have been claimed, roughly 40% of the run has probably sold, leaving about 2,400,000 tickets. That is an approximation — claim rates lag sales, and some prizes are never claimed at all — so treat the result as an estimate with a real error bar, not a precise figure.
The scratch card EV calculator does this arithmetic once you supply the published table.
Top prizes gone, game still on sale. This is the clearest signal. A game whose two $500,000 prizes have both been claimed but which is still in the racks is strictly worse than it was at launch — sometimes dramatically. The worked example shows a game dropping from 68% to 58% return on exactly this event.
A new game versus an old one at the same price. A freshly launched game has its full prize structure intact. All else equal it is the better buy, and this is the only "strategy" in instant games that is both real and free.
Unusually good remaining ratios. Occasionally a game's mid-tier prizes are disproportionately unclaimed. This raises EV genuinely — though almost never above the ticket price, for reasons below.
1. It almost never turns positive. Games launch at 63–74% return depending on price point (Minnesota's published figures). A favourable prizes-remaining position might lift a game a few points. Getting from 68% to over 100% would require nearly every large prize to remain in a nearly-sold-out run — which operators specifically avoid by ending games once top prizes are gone.
2. Tickets remaining is an estimate, and the estimate does the most work. Small errors in the denominator move the answer more than anything in the numerator. Be suspicious of any tool presenting a scratch EV to the cent.
3. Operators end games. Most lotteries close a game once its top prizes are claimed, precisely so the racks are not full of degraded tickets. The window in which a badly-degraded game is still on sale is usually short.
Prizes-remaining data will not make scratch cards profitable. It will let you:
That third one is the most valuable. Working through the arithmetic on a game you actually play — 68 cents back per dollar, before you account for buying the next one with the winnings — is more informative than any general warning about gambling.
See also: why the printed odds are not your odds and scratch card expected value, calculated properly.
Last verified: 2026-08-29