NBA player prop projections
Projects scoring, rebounding, and playmaking props with a rotation model that decides availability, who comes off the bench, and how minutes are split in every simulation. Live on real lines this season, opening night included, where every player is projected from last season.
A free account opens the board and the graded record. Choose NBA in the sport selector.
Today's free sample
The props the free plan opens on the 2026-11-16 NBA slate are among the props below, locked on this public page. A free account opens them on the board and the grid. The Model Room holds the full graded record. Only two of the free plan's three props on the 2026-11-16 NBA slate are still ahead, so the sample shows those and no others.
Grades, edges, probabilities and projections are not shown on this public page.Open today’s free sample with a free accountAlready signed in? The sample opens on the Signal Board.
6481 props listed here: 2 open with a free account, the rest in Pro
- AJ DybantsaWSH vs MEM · 3PM · 0.5 · PrizePicksOdds1 venueOver -251Under +167
- AJ DybantsaWSH vs MEM · 3PM · 1.5 · PrizePicksAvg odds3 books and marketsOver +179Under -244
- AJ DybantsaWSH vs MEM · 3PM · 1.5 · SleeperAvg odds3 books and marketsOver +179Under -244
The free plan opens three props per sport each slate. Pro opens every prop.
NBA markets and the graded record
The NBA standard-line record over the last 30 days is 50.9% across 49,808 decided lines. 21 of 21 markets have a graded record in that window.
- 49.4%Points48.3% to 50.5% at 95% · n=7,336
- 51.7%Rebounds50.5% to 52.9% at 95% · n=6,305
- 54.8%Assists53.3% to 56.3% at 95% · n=4,433
- 54.6%Three-Pointers Made53.2% to 56.0% at 95% · n=4,928
- 51.6%Steals47.2% to 56.0% at 95% · n=498
- 48.7%Blocks41.7% to 55.8% at 95% · n=189
- 50.8%Turnovers45.8% to 55.8% at 95% · n=378
- 50.1%Points + Rebounds + Assists48.9% to 51.3% at 95% · n=6,969
- 48.5%Points + Rebounds46.8% to 50.2% at 95% · n=3,378
- 50.8%Points + Assists48.9% to 52.7% at 95% · n=2,743
- 51.0%Rebounds + Assists49.2% to 52.8% at 95% · n=2,905
- 47.8%Fantasy Score45.5% to 50.1% at 95% · n=1,785
- 48.2%Field Goals Made45.2% to 51.2% at 95% · n=1,054
- 50.1%Field Goals Attempted47.4% to 52.8% at 95% · n=1,362
- 47.6%Three-Pointers Attempted44.5% to 50.7% at 95% · n=1,008
- 49.2%Two-Pointers Attempted46.3% to 52.2% at 95% · n=1,098
- 49.7%Two-Pointers Made46.1% to 53.4% at 95% · n=717
- 49.6%Free Throws Attempted46.2% to 53.0% at 95% · n=810
- 50.7%Free Throws Made46.9% to 54.5% at 95% · n=651
- 52.7%Offensive Rebounds47.9% to 57.5% at 95% · n=414
- 52.0%Defensive Rebounds47.5% to 56.5% at 95% · n=473
Rates are the standard-line record over the last 30 days: every decided standard line, recommended or not, wins and losses. The recommended cohort is reported as its own block in the Model Room and is not this number. Goblin and demon lines are over-only products and are not inside these numbers. Intervals are Wilson at the raw sample size, and because rows share slates and games they are a floor on the uncertainty.
What the NBA engine models
Plays out each game it projects possession by possession, thousands of times, with the rotation decided in each one
- A rotation model decides in every simulation who is available, who comes off the bench and how the minutes are split, so a player's range carries the nights his role changes
- When a star sits, his touches move to the teammates who handle the ball, in proportion to their role
- Blowouts, garbage time, foul trouble and rest on back to back nights are played out in the simulation, not added afterward
- League conditions such as pace, home court and overtime are measured from the current season's games, falling back to the previous season until enough current season evidence exists
- Eligible settled projections are checked against results, and the record shows the selected window and the engine version where one was recorded
Serving engine version nba-sim-1.2
What to know before you read it
The engine is built and has passed two audits. On opening night every player and team is projected from last season, and current season games take over as they are played. A player's first game of the season is shown as a projection, not a recommendation. The rotation model is confirmed on the 2025-26 season and has no live record yet, so every NBA grade is capped until that record builds. Availability reads published injury reports, projected lineups and injury news, and the board rebuilds in the hours before tip as starting lineups are confirmed. Graded nightly from final box scores, and a leg that depends on a platform rule we have not verified is left ungraded rather than settled.
No sportsbook consensus for NBA player props enters a grade or an edge. Market % stays empty by construction, so edges are model versus line. Where sportsbooks quote the same prop, their prices may be shown beside the line for comparison only, and they never set an edge or a grade.
- Simulated scoring runs noticeably wider than real games, which tighten when one team pulls ahead, so the chance of a very high or very low total can read too large
- Garbage time and rotation changes are set at the start of each simulation rather than by the simulated score, so how props in the same game move together is built into the design and not yet confirmed against real results
- The rest and schedule effects come from last season and are not yet confirmed on this season's games
- A player or team with no game this season is projected from last season, and current season games take over as they are played, so the first weeks rest on last season's role and a handful of new games
- Availability reads published injury reports and injury news before each build, so news that breaks after the last build is not reflected, and a player no source spoke for is shown as unknown, never healthy
- The rotation model is confirmed on a past season only and grades are capped until its live record builds
NBA research
- Three Pointers Made Props in the NBAA threes made line of 2.5 looks like a question about shooting. It is mostly a question about attempts, asked of a small count distribution where the average sits in an awkward place and a single miss moves the result. Streaks make it feel more predictable than it is.
- NBA Rebound and Assist PropsA rebound needs somebody to miss. An assist needs somebody else to score. Both props depend on events the player does not control, which is why they are best researched as opportunity first and conversion second, with pace setting how many chances there are at all.
- Reading the First Weeks of an NBA SeasonThe first two weeks of an NBA season produce the most confident takes and the least evidence. Last season is still most of what you know, but not all of it transfers, and some early numbers deserve weight long before others do. The skill is knowing which.
- Usage When an NBA Star SitsA star's absence frees two things at once: his minutes and his share of the offense. They go to different people. The minutes follow the depth chart, the shots follow shot creation, and the data most people reach for to measure the shift is thinner than it looks.
- Minutes, Blowouts, and NBA Player PropsEvery NBA counting prop sits on top of a minutes estimate, and most minutes estimates are a single number pretending to be certain. Treat playing time as a range with a lopsided tail, read the spread as a warning about that tail, and split the minutes log by how close the game was.
- How to Research NBA Player PropsCounting stats in basketball are a product of two numbers multiplied together, and almost every research mistake comes from studying the second one while ignoring the first. Start with time on the floor, then ask what share of the offense runs through the player while he is on it.
Read the full NBA board
A free account opens the Signal Board, the Prop Grid and the Model Room. Research and analysis only, never a bet placed for you.
The Slate
One weekly note: what published, what the graded record shows, what changed. No picks, and nothing is sent until you confirm your address.