Learn
The Hot Hand Check: an NBA player prop hit rate table, then the price
An NBA player prop hit rate over ten games comes with a wide honest range, and the Hot Hand Check puts that range next to the price before the streak gets a vote. Streaks can be real, so the check is not about ignoring them: it is sample first, then price, and only then the streak.
Ten straight games of overs on a points, rebounds and assists line feels like information. Some of it is. The Hot Hand Check is a way to use a streak without overpaying for it, and the range table below is the part to keep next to your prop research.
The order from the video:
- Sample first. Put the record into its honest range.
- Then price. Find the fair chance for the line with the margin removed.
- Only then the streak. Let the streak nudge your estimate, never replace the price check.
Who this is for
This is for bettors who research NBA player props and see a lot of last-five and last-ten hit rates. You need two prices for the prop, a calculator that removes the margin, and the player's longer-run numbers. The tables here work for any over-under prop, whether it is points alone, rebounds, assists or a combined line, as long as you read the record as a sample and not as a promise.
Streaks can be real
For decades the standard view was that the hot hand is a fallacy. Miller and Sanjurjo showed in Econometrica in 2018 that the classic measure of streak shooting has a subtle but substantial bias, and that correcting for it reverses the original conclusion. So a streak can carry some signal. That's a reason to measure it properly, not a reason to bet it at any price.
Step 1. Sample first
Ten games is about 12% of an 82-game season. The table gives a 95% Wilson interval for each possible ten-game record: the range of true over rates that would plausibly produce it.
| Overs in last 10 | Hit rate | Honest range (95% Wilson) |
|---|---|---|
| 0 | 0% | 0% to 28% |
| 1 | 10% | 2% to 40% |
| 2 | 20% | 6% to 51% |
| 3 | 30% | 11% to 60% |
| 4 | 40% | 17% to 69% |
| 5 | 50% | 24% to 76% |
| 6 | 60% | 31% to 83% |
| 7 | 70% | 40% to 89% |
| 8 | 80% | 49% to 94% |
| 9 | 90% | 60% to 98% |
| 10 | 100% | 72% to 100% |
And for twenty games, if your tool shows them:
| Overs in last 20 | Hit rate | Honest range (95% Wilson) |
|---|---|---|
| 10 | 50% | 30% to 70% |
| 12 | 60% | 39% to 78% |
| 14 | 70% | 48% to 86% |
| 16 | 80% | 58% to 92% |
Read the 6-of-10 row. A player who went over six times in ten could plausibly be a 35% over or an 80% over. On its own, that record can't tell those apart, which is the whole point of the interval. Twenty games narrows the range, but not by as much as most bettors expect.
Step 2. Then price
Now find the fair chance for the line. Convert both sides to implied chances, add them, and divide each by the total. For lopsided prices, the power method did better than simple normalization in a 2017 comparison by Clarke, Kovalchik and Ingram. Our no-vig calculator offers both.
| PRA 38.5 (hypothetical, demo odds) | Over | Under | Total |
|---|---|---|---|
| Price | -120 | +100 | |
| Implied chance | 54.55% | 50.00% | 104.55% |
| Fair chance | 52.17% | 47.83% | 100.00% |
| Fair odds | about -109 | about +109 |
Then compare prices across books:
| Over price (demo odds) | Expected value per $100 at a fair 52.17% |
|---|---|
| -120 | about -$4.35 |
| -110 | about -$0.40 |
| -105 | about +$1.86 |
| +100 | about +$4.35 |
The same over that loses about $4.35 per $100 at -120 is worth about $1.86 at -105. The streak didn't change between those rows. The price did. You can check any row in the expected value calculator.
Step 3. Only then the streak
Now let the streak have its say, in proportion:
- Start from a long-run estimate of the player's over rate on this kind of line (season and last season together).
- Nudge it for real changes: a bigger role, more minutes, a teammate out. These are reasons a streak might be real.
- Check rest. Yang and colleagues found paint and three-point scoring were significantly lower in back-to-back games, so a hot run can cool on the second night.
- Compare your estimate with the fair chance. Bet only if yours is clearly higher, and only at a price better than fair.
The one-page check
| Field | Your entry |
|---|---|
| Player, line | |
| Last 10 record and honest range | |
| Both prices, fair chance for the over | |
| Your long-run estimate, with role and rest adjustments | |
| Best over price at any book, expected value | |
| Bet? | only if your estimate beats the fair chance and the price beats fair odds |
What to watch out for
- Counting the streak twice. If your long-run estimate already includes the last ten games, don't add a hot-hand bonus on top.
- Line moves. A streak often moves the line itself, so the bargain may already be gone.
- Small rate changes in big numbers. A PRA line moves with minutes far more than with shooting luck.
How to tell it worked
Log each prop with the honest range, the fair chance, the price you took and the closing price. After a month, the share of your prices that beat the close is the real test. A hot-hand bet that beat the close and lost was still a good bet. The Quant+ player prop research page shows the last ten games, the best price and the fair price on one row.
21+. For informational purposes only. Nothing is guaranteed. If gambling stops being fun, call 1-800-GAMBLER.
Questions people ask
Is the hot hand real?
The famous 1985 finding that it was a fallacy had a measurement bias. Miller and Sanjurjo (Econometrica, 2018) showed that correcting it reverses the original conclusion, so streaks can carry some information.
What is the honest range for 6 overs in 10 games?
A 95% Wilson interval for 6 of 10 runs from about 31% to 83%. That is too wide to tell a hot player from an average one on its own.
How do I get a fair price for a PRA prop?
Convert both sides to implied chances, add them, and divide each side by the total. In a hypothetical 38.5 line at -120 / +100 the fair over is about 52.2%.
Does Quant+ show last-ten hit rates?
Yes, the props board shows the last ten games next to the best price, the fair price and the edge. It does not place bets or promise results. 21+.
Sources
- Surprised by the Hot Hand Fallacy? A Truth in the Law of Small Numbers, Econometrica 86(6) (Miller, Sanjurjo)
- Influence of Match Congestion on Performances in the National Basketball Association, Frontiers in Psychology (Yang et al.)
- Adjusting bookmaker's odds to allow for overround, American Journal of Sports Science (Clarke, Kovalchik, Ingram)
- 2025-26 NBA season, Wikipedia
- Binomial proportion confidence interval, Wikipedia
