NBA Player Prop Hit Rates
What our points, rebounds, assists and double-double props actually hit last season, graded in public — and how to find value in NBA player props using the same method.
Our NBA player prop hit rates, graded
Most NBA props content tells you what to look for. This page starts with what happened. Every play the model published above its threshold from March 12, 2026 to June 13, 2026 (79 game days, playoffs included) was settled against the official box score the next morning, and the counters below are the result. The break-even at -110 is 52.4%.
| Market | What counts as a play | Record | Hit rate |
|---|---|---|---|
| Rebounds † | model ≥ 65% vs its own line | 203-90 | 69.3% |
| Double-double | top pick, model ≥ 60% | 93-53 | 63.7% |
| Assists | top pick, model ≥ 65% | 450-337 | 57.2% |
| Triple-double | top pick, model ≥ 30% | 11-10 | 52.4% |
| Points | top two, model ≥ 65% | 462-451 | 50.6% |
Green clears -110; amber does not. Source: the same counters as the performance page, rebuilt from the archived pick files on September 20, 2026. † Rebounds is graded against a line the model sets from the player's own 8-game average (our odds feed carries no rebound lines; 260 of the 293 plays were unders at the 2.5 floor), so it is not a market hit rate and is shown grey.
Read it honestly. Double-doubles are where the model earns its keep: a role market, driven by minutes rather than whether shots fall, and it cleared the price by a wide margin over 146 plays. Assists at 57.2% beats -110 comfortably over 787 plays and is the board we publish free. The rebounds row is the model against its own line and says nothing about a price. Points at the play tier hit 50.6% of 913 — a coin flip that loses to the vig. Points is the most heavily bet prop in the sport, the line is the sharpest, and our model has the least to add there. We keep publishing the points projection because the number is useful context; we do not tell anyone to bet it blind.
The team markets, for the same window: moneyline 156-41 (79.2%) at model ≥ 60%, spread 155-52 (74.9%) at projected margin ≥ 5, totals 58-46 (55.8%) in the 60–70% band. The moneyline number is a hit rate, not a return: most of those favourites were priced well past -200, so the question on any given night is the price, not the percentage.
How to find value in NBA player props
Value is a probability that is higher than the price implies. At -110 the book is saying 52.4%; if you have a defensible reason to think the true chance is 58%, that is a bet, and if you don't, it isn't — however good the story sounds. Everything below is about building that probability from things the line is slow to price.
Minutes first, everything else second
A starter at 36 minutes has a quarter more opportunities than one at 28. Lines are anchored on season averages and react to rotation changes over three or four games; the box scores react the same night. Track the last five games' minutes, not the season figure, and treat a coach quietly extending a role as the single most exploitable situation in the market.
Usage moves when a teammate sits
When a team's second scorer is out, the primary option's usage jumps and the points line rarely adjusts fully. When the starting point guard sits, the backup's assist line is set to his season average (say 4.5) while he is about to play 34 minutes as the only ball-handler. Check the injury report in the 90 minutes before tip-off; the lines you see at noon were built on a different lineup.
Pace and the game total
A 230-point total is a fast game with more possessions for everyone; a 210-point total is a grind. Books adjust for pace but not always enough in the extremes — two fast teams meeting, or a fast team on the second night of a back-to-back. Rebounds also depend on the shot profile: long three-point misses produce long rebounds, which favour guards and wings over centres.
Blowout risk is why overs fail
A 12-point favourite's starters may sit the whole fourth quarter. A prop priced on 34 minutes becomes a prop played in 26, and every over on that side dies together. This is the number one reason overs lose and the reason our model weights projected competitiveness.
Back-to-backs, rest and home splits
Minutes restrictions on the second night of a back-to-back suppress every line, most for older players and big men; some books still under-adjust. Two days of rest tends to mean full starter minutes. Home/away splits of five points or more are real for some players and invisible in a line set near the overall average.
Shop the line and time it
One book at 22.5 and another at 23.5 is a full point of edge for free. Openers are hit by sharp money within minutes; late injury news creates value again just before tip. Check when lines open and again 30 minutes before the game — and never parlay more than two of these: a four-leg -110 parlay holds about 17% for the book.
How our model produces the numbers
Points and assists are XGBoost regressors retrained daily on the current season's player box scores. The features are the things above, measured rather than eyeballed: a rolling ten-game window of minutes, usage and production, the opponent's pace and defensive profile, home/away, and rest. The regressor projects the stat; the over/under probability then comes from the player's own recent eight-game distribution around that projection, so a consistent scorer at 24 points gets a sharper probability than a streaky one at the same average. Players under 20 minutes are excluded. A play is the top of the board at 65% or better.
Rebounds are different, and the difference matters: our odds feed carries no rebound lines, so the board sets its own from the player's eight-game average and grades against that. It is a projection to hold against your book's line, not a graded market. Double-double and triple-double probabilities come from the player's historical rate, with plays at 60% and 30% respectively.
Two things the model does not do. The regressor never takes the betting line as an input — the projection is made blind, and the line only enters afterwards as the number the projection is measured against, which is what makes the graded record an honest test rather than a copy of the book. And it does not know a player is out until the report says so, which is why the last 90 minutes before tip-off belong to you, not the model.
The markets, briefly
Points
The most popular and most efficiently priced market. Edges come from lineup changes and minutes, not from the season average, which the book already has.
Rebounds
Position, minutes, matchup and the opponent's shot profile. Sensitive to blowouts. Our board projects the number without a book line; use it against yours.
Assists
Pace, teammate shooting and who else handles the ball. A missing shooter lowers the primary passer's line even as his usage rises. Free board, graded 57.2%.
Double-doubles and triple-doubles
Role props: a player averaging 18 and 9.5 is one rebound from a double-double most nights, and the price often treats it as a coin flip. Triple-doubles are offered for a handful of players; a 30% probability at +250 is value, at +150 it is not.
Combos (PRA, PR, PA)
Aggregating three stats lowers variance. Good for well-rounded players, poor for one-dimensional scorers. We do not publish combo props; build them from the individual projections.
The NBA prop boards
Assists projections for every player, free, with the last ten games against the line. Points and double-doubles: the top five free, the full boards for subscribers. Rebounds: a projection to hold against your book's line. First board of the season: opening night, October 20. Every play is graded the next morning and added to the record above.
Read the methods
- How to predict NBA player points props — the z-score + regression method, and why points graded 50.6%.
- NBA double-double props: how the historical-rate model works — a calibrated rate, and the price each tier needs.
- How an NBA moneyline model reaches 79% — the 23 features, the bands, the playoff drop, and what the number does not cover.
Frequently asked questions
What hit rate do you need to profit on NBA player props?
At the standard -110 price you need 52.4% to break even; at -115 it is 53.5%. A hit rate is only meaningful against the price you paid, which is why we publish ours alongside the play threshold that produced it.
What are your NBA player prop hit rates?
Graded from March 12 to June 13, 2026 (79 days, playoffs included): double-doubles 63.7% of 146 plays, assists 57.2% of 787, triple-doubles 52.4% of 21, points 50.6% of 913. Rebounds went 69.3% of 293, but against a line the model sets from the player's own average — no book line is involved, so it is not comparable. The points model at the play tier did not beat -110; we say so on the page.
How do you grade a prop?
Every published play above the threshold is settled against the official box score the next morning, by player and date, and the counters are kept on the site. Nothing is removed after the fact. The full record was rebuilt from the archived pick files on 2026-09-20 so every number is reproducible.
What is the best NBA player prop market to bet?
By our graded record, double-doubles — a role market the books price lazily — and assists, which cleared -110 over 787 plays. Points is the most popular market and the most efficiently priced; it is where our model has the least to add. Our rebounds number is graded against a self-set line and does not answer this question.
When do the NBA boards start?
The first board of the 2026-27 season is published on opening night, October 20, 2026. Assists projections are free; points, rebounds, double-double and triple-double boards are for subscribers. Last season's record stays on this page.
The week ahead, once a week
Last week's graded record — wins and losses — the plays for the week ahead, and a subscriber discount code. One email, no daily noise, unsubscribe in a click.
Related
Responsible Gambling Disclaimer: Sports betting involves risk and should be done responsibly. Never bet more than you can afford to lose. Prediction Engine provides data-driven analysis for educational and informational purposes only. Hit rates are historical and are not a guarantee of future results. If you or someone you know has a gambling problem, call 1-800-GAMBLER or visit ncpgambling.org. Must be 21+ and located in a jurisdiction where sports betting is legal.