Scratch cards and instants
Scratch cards genuinely beat draw games on every static measure. They lose decisively on the one measure that determines what you actually spend.
On the numbers, this is not close. Instant games win on return, on frequency of winning, and on top-prize odds. And they are, for most people, the more expensive product.
Return to player. Minnesota's published figures run from 63% at $1 to 74% at $50. Draw games typically return around half of sales — the sourced comparison is in return to player by game and the payout ratio league table.
Frequency of any win. A typical scratch game wins something on roughly 1 ticket in 3 to 1 in 5. US Powerball pays any prize about 1 in 24.9 — and its most common win is $4 on a $2 ticket.
Top-prize odds. A $500,000 scratch top prize might be 1 in 2,000,000. Powerball's jackpot is 1 in 292,201,338 — about 146 times longer. Scratch top prizes are smaller by a similar factor, which is exactly why the odds are shorter.
Computability. You can work out an instant game's current expected value from published prizes-remaining data (how). No such thing exists for a draw game.
On every static measure, instant games are the better product. So why do they have a worse reputation among people who study this?
Expected loss is not a percentage. It is percentage × amount staked, and staking is limited by how fast you can play.
Put numbers on it. Say you play $20 a week either way:
| Draw game | Scratch cards | |
|---|---|---|
| Return | ~50% | ~68% ($5 tickets) |
| Weekly spend | $20 | $20 |
| Expected weekly loss | $10 | $6.40 |
At equal spend, scratch cards are better — by $3.60 a week. That is real, and it is the honest case for them.
But equal spend is the assumption that fails. The scratch player's $20 is not a fixed subscription; it is a stack of tickets, roughly a third of which pay out, and the payouts are handed over at the counter where the next ticket is sold.
This is the mechanism that matters, and it is specific to instant games.
Buy a $5 ticket, win $5, buy another. You have "not spent anything" and you have staked $10 against a 68% return. Do it again and it is $15. Each cycle takes seconds and feels free, because the money never left your pocket in a form you noticed.
Running $20 through a 68% game repeatedly is not a 68% return on $20. It is 68% of 68% of 68%… — the expected value of your original stake decays geometrically with each cycle. After four cycles, $20 has an expected surviving value of 20 × 0.68⁴ ≈ $4.28.
A draw game cannot do this to you. You buy on Tuesday, you find out on Wednesday, and your winnings arrive as an event with a gap after it — a gap in which a decision can happen.
That structural difference is why the product with the better percentage is frequently the worse habit, and it is reinforced deliberately by design choices covered in why the near-miss is deliberate.
| Draw game | Scratch card | |
|---|---|---|
| Return to player | ~50% | 63–74% |
| Odds of any prize | ~1 in 25 | ~1 in 3–5 |
| Top prize odds | 1 in 292m | 1 in ~2m |
| Top prize size | hundreds of millions | usually under $1m |
| Current EV computable? | No | Yes, from prizes remaining |
| Plays per hour | 0 | dozens |
| Feedback delay | days | seconds |
If you play a fixed amount and stop, scratch cards are the better buy. That is a genuine, sourced conclusion, and it is the opposite of what most anti-lottery writing says.
If your spend is determined by how the session is going, the draw game is safer — not because it is a better product, but because it makes it impossible to have a session.
Neither is an investment: both have negative expected value, and no scratch game reaches break-even (positive expected value in a lottery). The useful question is not which loses less per dollar, but which one lets you decide in advance how many dollars there will be.
Last verified: 2026-08-29