Basketball Possessions and Pace Calculator
Basketball Possessions and Pace Calculator
Estimated possessions and pace in basketball, computed two published ways at once so you can see how far apart they are. No league counts possessions, and the free-throw coefficient everything rests on is an empirical fit.
Two published possession estimates, and the gap between them
Estimate, not a countThe NBA’s 2023-24 league average per game: 88.9 field goal attempts, 42.2 made, 21.7 free throw attempts, 10.6 offensive rebounds, 13.6 turnovers, 33.0 opponent defensive rebounds
Two published formulas, and the coefficient they disagree about
Longer form, one team = FGA + 0.4 × FTA − 1.07 × (ORB ÷ (ORB + Opp DRB)) × (FGA − FG) + TOV
Pace = possessions × 48 ÷ regulation game length in minutes
- nobody counts possessions
- the NBA’s own stats glossary defines possessions in words — “The number of possessions played by a player or team” — and gives no formula, and defines pace as “The number of possessions per 48 minutes” with no formula either. A 2023 paper on the four factors states flatly that “the exact number of possessions can not be deduced” from a box score
- the simple estimate
- Dean Oliver’s approximation, published in Basketball on Paper (Potomac Books, 2004), usually written POSS = FGA + TOV − ORB + μ × FTA. This page uses μ = 0.44, the value in the NBA’s own glossary; the NBA’s points-per-possession page uses 0.436, and a variant scaling the whole estimate by 0.96 circulates on the APBRmetrics forum, attributed there to the NBA, unverified in that thread and not computed here
- the longer form, and why the two disagree
- Basketball-Reference’s team formula uses 0.4 rather than 0.44 — worth about 0.9 of a possession here — and estimates the offensive rebounds as 1.07 times the team’s share of available rebounds times its missed shots, removing 12.15 against the 10.6 recorded, which is most of the gap. The published version then AVERAGES that with the same calculation from the opponent’s line, needing the opponent’s whole box score; this page computes the one-team half, a third variant, and names it as one
Worked example
The NBA's 2023-24 league average per game: 88.9 field goal attempts, 42.2 made, 21.7 free throw attempts, 10.6 offensive rebounds, 13.6 turnovers, 33.0 opponent defensive rebounds
Simple estimate: 88.9 + 0.44 × 21.7 − 10.6 + 13.6 = 101.4 possessions
The longer form's rebound share: 10.6 ÷ 43.6 = 0.243, so 1.07 × 0.243 × (88.9 − 42.2) = 12.15 estimated offensive rebounds, more than the 10.6 recorded. That gives 88.9 + 0.4 × 21.7 − 12.15 + 13.6 = 99.0 possessions
Divergence 2.42, or 2.4%, on identical inputs. Downstream, at 114.2 points the two give offensive ratings of 112.6 and 115.3 — 2.75 points per 100 possessions of pure definitional difference
The coefficient alone: 0.40 gives 100.58 and 0.475 gives 102.21, a spread of 1.63. Pace: these are per-game figures from a 48-minute game, so pace equals the possession figure; switch the selector to a 40-minute game and the same possessions become a pace of 121.7
Five published values for one coefficient, on the same line
| Coefficient | Where it was read | Possessions | Offensive rating at 114.2 points |
|---|---|---|---|
| 0.40 | Oliver, Basketball on Paper, 2004 | 100.58 | 113.54 |
| 0.42 | arXiv:2305.13032, refitted on 2022-23 data | 101.01 | 113.06 |
| 0.436 | The NBA’s own points-per-possession page | 101.36 | 112.67 |
| 0.44 | The NBA stats glossary and Basketball-Reference’s | 101.45 | 112.57 |
| 0.475 | Zimmermann (2016), for NCAA basketball | 102.21 | 111.73 |
Two published formulas on identical inputs
| Simple estimate | Longer form, one team | |
|---|---|---|
| Published by | Oliver, Basketball on Paper (2004), with the NBA glossary’s coefficient | Basketball-Reference’s glossary |
| Free throw coefficient | 0.44 | 0.4 |
| Offensive rebounds | Subtracted as recorded: 10.6 | Estimated at 1.07 times the rebound share times the misses: 12.15 |
| Needs the opponent’s box score | No | Yes — the published version needs the whole opponent line, because it averages both teams’ estimates |
| Possessions on this page’s default line | 101.4 | 99.0 |
| Measured error over 2016-17, counted from play-by-play | Not measured in the sources read | Off by 2,189 over the season, 1.93 a game, worst team 158 |
The number everything else divides by is an estimate
This page is about basketball, the five-a-side game played to a basket ten feet above the floor. It has no single rule-maker: the Fédération Internationale de Basketball (FIBA) publishes the Official Basketball Rules used by most of the world, the NBA publishes its own rulebook for its own competition, and the NCAA publishes separate rules for United States college play. Where a number on this page is one of theirs rather than plain arithmetic, the body and the edition are named beside it.
A possession count is an estimate, not a count. No league publishes one alongside its box score: the NBA’s own stats glossary defines possessions in words — “The number of possessions played by a player or team” — and prints no formula for them at all, and defines pace as “The number of possessions per 48 minutes” with no formula either. Every possession figure you have read was reconstructed from field goal attempts, free throw attempts, offensive rebounds and turnovers, and the published reconstructions do not agree with each other. The reconstruction is simple in outline: a possession ends in a shot, a turnover or a trip to the free throw line, so add those and subtract the offensive rebounds, which mean the possession carried on rather than restarting. That is Dean Oliver’s approximation, published in Basketball on Paper in 2004. The part that is not right is the free throws, which is why each attempt counts as a fraction of a possession-ending event.
The 0.44 is an empirical fit, not a rule of basketball. It stands for the proportion of free throws that end a possession, which is not 1 — a trip to the line is usually two shots, the and-one adds a throw to a possession that already ended in a basket, and a technical foul ends nobody’s. The value is attributed to Kubatko and colleagues (2007) fitting 2003-06 data, and it is the figure the NBA’s own glossary uses. Five published values are in the first table below, from Oliver’s 0.40 to the 0.475 used for college basketball, and on this page’s opening line that range moves the answer by 1.63 possessions a game. Worse, Basketball-Reference uses 0.44 in true shooting percentage, usage rate and turnover percentage and 0.4 in its own possession formula, on one glossary page — two numbers for one physical quantity, printed here rather than quietly reconciled.
The formulas also differ structurally, which is why this page computes both and prints the gap. The simple estimate subtracts the offensive rebounds the team took; the longer form estimates them at 1.07 times the team’s share of available rebounds times its missed shots, then averages the result with the same calculation from the opponent’s line. On identical inputs the two come out 2.42 possessions apart, about 2.4%, and because every rating and per-100 figure divides by this number the same 114.2 points become an offensive rating of either 112.6 or 115.3 — 2.75 points per 100 of pure definitional difference, larger than many of the gaps people draw conclusions from. Never compare a per-possession figure from one source with one from another, and say which estimator you used. On how good they are there is one direct measurement, in the second table: Squared Statistics counted 2016-17 possessions from play-by-play and found the long form out by 1.93 a game and the NBA’s own estimator by 6.29, with a refit giving 0.4328 and 1.2228 rather than 0.4 and 1.07 — so even the better formula’s constants belong to one era. To turn a possession figure into points per 100 use the offensive and defensive rating calculator.
Frequently asked questions
How do you calculate possessions in basketball?
You estimate them. The common simple form is field goal attempts plus 0.44 times free throw attempts, minus offensive rebounds, plus turnovers. For the NBA’s 2023-24 league average: 88.9 + 9.55 − 10.6 + 13.6 = 101.4 a game. It is an estimate because no league publishes a count and a box score does not contain one.
Why is an offensive rebound subtracted?
Because it continues a possession rather than starting one. The NBA’s own stats glossary says “an Offensive Rebound does not create another possession, it simply makes the existing possession longer”. Counting it as new would credit a team with extra possessions for its own missed shots.
Where does the 0.44 free throw coefficient come from?
From fitting data, not a rule. It estimates the share of free throws that end a possession, and 0.44 is attributed to Kubatko and colleagues on 2003-06 NBA data. The NBA’s own page uses 0.436 and says the figure “may change from year to year”; a 2023 refit gives 0.42; college basketball uses 0.475; Oliver used 0.40.
Why do two sites give different possession numbers for the same game?
Because they use different published formulas. Here the simple estimate gives 101.4 and Basketball-Reference’s longer form 99.0 — a 2.4% gap from one box score, because the long form uses a different coefficient and estimates offensive rebounds rather than subtracting the recorded ones. Neither is wrong; they are different definitions.
What is pace, and why is it per 48 minutes?
Pace is possessions per 48 minutes of team play, and 48 because that is the length of the NBA’s regulation game. A FIBA game is 40 minutes, so its possessions must be scaled by 48/40 before the figures compare. Both period structures are quoted with their article numbers on this site’s three-point line and shot clock page.
Related calculators
References
- National Basketball Association. Stats Glossary, nba.com, read 12 Oct 2026. The source for this site’s true shooting percentage, effective field goal percentage and offensive, defensive and net rating formulas, each quoted verbatim on the page that uses it. Possessions and pace are defined in WORDS with NO FORMULA at all. Usage percentage is “(FGA + Possession Ending FTA + TO) / POSS” and turnover ratio “(TO * 100) / (POSS)” — neither the 0.44 form.
- Sports Reference LLC. Basketball-Reference.com Glossary, read 12 Oct 2026. True shooting attempts “FGA + 0.44 * FTA”, eFG% “(FG + 0.5 * 3P) / FGA”, TOV% “100 * TOV / (FGA + 0.44 * FTA + TOV)”, Pace “48 * ((Tm Poss + Opp Poss) / (2 * (Tm MP / 5)))”. Its team Possessions entry instead uses 0.4 on free throw attempts and 1.07 on an estimated offensive-rebound term, and averages both teams’ estimates — so one page carries 0.4 and 0.44 for the same quantity. Offensive and defensive rating carry no formula here.
- “Dean Oliver’s Four Factors Revisited”, arXiv:2305.13032, read 12 Oct 2026. States Oliver’s approximation as POSS = FGA + TOV − ORB + μ·FTA, with μ “the proportion of free throws that end the possession of the ball”; attributes μ = 0.44 to Kubatko et al. (2007) on 2003-06 data; refits μ = 0.42 for 2022-23, “no longer accurate in 2023”; cites Zimmermann (2016) using 0.475 for the NCAA; and states that “the exact number of possessions can not be deduced” from a box score. Quotes Oliver’s four-factor weights of 40/25/20/15%, calling them “somewhat ad hoc”. Cites Oliver D, Basketball on Paper, Potomac Books, 2004. Its figures are image-only.
- Squared Statistics. “Analyzing NBA Possession Models.” squared2020.com, 10 July 2017, read 12 Oct 2026. Quotes the NBA’s own estimator with its .436 coefficient and the NBA’s stated basis for it — “approximately 43.6% of all free throws are possession-ending”, which “may change from year to year”. Counts 2016-17 possessions from play-by-play and scores both formulas against the count; the figures and the refitted 0.4328 and 1.2228 are quoted on this site’s possessions page. Its per-team tables are images, so no row here comes from one.
- APBRmetrics forum, “Why are possessions estimated using crude formulae, instead of being properly and precisely counted?” Cited only to record that a further variant circulates — a poster’s stated belief that the NBA uses “Pos = .96 * (FGA + (.44*FTA) – OR + TO)” — and that it is unverified in the thread, which contains no measured accuracy figure for any estimator. A forum is the weakest kind of source and is used here for nothing else. Read 12 Oct 2026.
- Sports Reference LLC. 2023-24 NBA Season Summary, read 12 Oct 2026. The Per Game Stats table’s League Average row, used as this page’s default inputs so no worked example rests on an invented box score: FG 42.2, FGA 88.9, 3P 12.8, FTA 21.7, ORB 10.6, DRB 33.0, TOV 13.6, PTS 114.2. Its own Advanced table was absent from the fetched text, so every Pace, rating or percentage figure here is this site’s own arithmetic on that row.