Scratch cards and instants
Buying a whole roll guarantees you winning tickets. It does not guarantee — and on the published payout percentages cannot deliver — a profit.
Scratch cards are distributed to retailers in packs, and it is true that a pack contains a controlled, non-random number of winners. Every so often someone concludes that buying a whole pack is therefore a way to win. The arithmetic disposes of this quickly, and the reason it fails is instructive.
Instant games are not printed randomly and then shuffled. Winners are distributed through the print run under controls that ensure prizes are spread reasonably evenly across packs and geography — so that every region gets winners, no retailer gets a pack of pure losers, and nobody can predict prize position from purchase history.
A consequence is that a full pack reliably contains roughly its share of the game's low-tier winners. That is a real property, and it is what "guaranteed winners" refers to.
Take the $5 game used throughout this cluster (68% return, matching Minnesota's published $5 figure), and a pack of 100 tickets:
You are guaranteed winners. You are guaranteed, in expectation, to lose $160 getting them. The guarantee and the loss are not in tension — the guarantee is about the count of winning tickets, and the loss is about their value, exactly the confusion dissected in why the printed odds are not your odds.
In the worked prize structure, 84% of all winning tickets pay $5 — the ticket price. A pack full of guaranteed winners is largely a pack full of refunds.
1. Big prizes are spread across packs, not within them. The whole point of the distribution controls is that the two $500,000 prizes in a 4,000,000-ticket run are somewhere in the run — not one per pack. Your pack of 100 has a 100-in-4,000,000 chance of containing a top prize: 1 in 40,000. Buying the pack buys the low tiers reliably and the top tier barely at all.
2. You lose the one real edge available. The only genuine improvement in instant games is choosing a game whose top prizes are still unclaimed (prizes remaining). Committing $500 to one pack of one game is the opposite of selective.
3. Variance is lower, which is not a benefit here. Buying 100 tickets from a 68% game brings your realised return much closer to $340 than a single ticket does. Reduced variance around a negative mean means you are more certain of losing about $160. If your reason for playing is the small chance of a large outcome, a pack removes exactly that.
Buying a whole pack would be profitable if a pack were guaranteed to contain more prize money than it cost. That would require a return above 100%, which no game has — published payout percentages run 63% to 74%. Operators set these percentages at design time and audit against them; a pack that reliably paid out more than it cost would be an accounting error rather than a strategy.
The genuine historical exploits all worked differently: Cash WinFall exploited a published roll-down rule that redistributed money into short-odds divisions; the 1729 Paris lottery was mispriced; the 1992 Irish syndicate bought a small combination space outright. None of them involved buying more of a normally-priced product.
The same work it does in guaranteed-win wheels for draw games: making a true, narrow, conditional statement sound like a promise about profit.
This pack contains winning tickets — true. Therefore buying it is profitable — does not follow, and is arithmetically false at any payout percentage below 100%.
If you want the strongest honest position on instant games, it is in which price point has the best return: given a fixed budget, buy the highest price point that budget covers. That is worth about eleven percentage points. Buying by the pack is worth nothing at all.
Last verified: 2026-08-29