Points per possession: why efficiency beats volume
Every basketball game is a fixed number of trips down the floor. Learn how points per possession works, how possessions are counted, and why scoring efficiency matters more than scoring totals.
Two players can both average 25 points a game and be worth very different amounts to their teams. One gets there on 17 shots, a handful of free throws and almost no turnovers. The other needs 25 shots, wastes a few possessions a night on giveaways, and leaves his team a little worse off every time the ball goes up. The box score line reads the same, and the scoreboard tells a different story. The difference lives in a unit that rarely appears in the newspaper box score: the possession.
A possession is one trip with the ball. A team gets it, tries to score, and the trip ends when the other team gets the ball back or a made field goal or the last of a set of free throws changes who has it. Because the two teams trade possessions in a game, they get nearly the same number of them. That symmetry is the quiet foundation of modern basketball analysis. If both teams get roughly 100 trips, the winner is the team that squeezes more points out of each one. Points per possession, abbreviated PPP, simply says how well a team or a player turned trips into points.
This article walks through what a possession is, how analysts count it when the official scorer does not, the shot-by-shot arithmetic that makes efficiency matter, and the places where the idea gets messy. It also explains why coaches who understood this long before the spreadsheets arrived, such as Dean Smith at North Carolina, were already running their offenses on the same logic.
The possession as the unit of the game
Basketball is unusual among team sports because it has no innings, no downs and no fixed number of at-bats. A game ends when a clock runs out. That makes it easy to confuse volume with success: the team that scores more wins, so the player who scores the most seems to be the most valuable. But the clock is not what limits scoring. Possessions are.
Think about a lopsided example. A team that plays at a frantic tempo and scores 120 points in a game may be worse on offense than a team that scores 100 in a methodical game. If the first team had 120 possessions and the second had 90, then the fast team scored one point per trip and the slow team scored more than 1.1. The raw totals reward speed. The per-possession view strips speed out and asks the cleaner question: how good was the offense when it had the ball?
This is why the league's own best-known efficiency numbers are expressed per 100 possessions. Offensive and defensive rating is nothing more than points per possession multiplied by 100 so that the figure reads as a round number. A team that scores 1.12 points per possession has an offensive rating of 112. The conversion exists because 1.12 looks like a decimal and 112 looks like a score, and fans find the second easier to hold in their heads.
There is a second reason to count in possessions: it forces you to notice that every possession has an owner and an outcome. A turnover is not just a lost ball. It is a trip that produced zero points, and in the same way a missed shot that the other team rebounds is a trip that produced zero points. Seen that way, basketball becomes a game of not wasting trips.
How possessions are counted
Unlike points, possessions are not printed in the standard box score. The official scorer does not tick one off each time the ball changes hands. For most of the league's history, nobody had the play-by-play detail to count them directly, so analysts estimated them from the numbers that were recorded: field-goal attempts, free-throw attempts, offensive rebounds and turnovers.
The most common estimate is easy to state. A team's possessions are roughly its field-goal attempts, plus a fraction of its free-throw attempts, minus its offensive rebounds, plus its turnovers. The logic is as follows.
- Field-goal attempts are the starting point. Each shot is, roughly, the end of one possession.
- Offensive rebounds are subtracted because a missed shot that the shooting team rebounds does not end the possession. The team shoots again, so two attempts happened inside a single trip.
- Turnovers are added because a turnover ends a possession without a shot ever being attempted.
- Free throws are the awkward part. A trip to the line can end a possession, but not every free throw does, because some are the first of a pair and some come after a made basket on an and-one.
That last bullet is where the famous 0.44 comes from. Over a large number of games, about 44 percent of free-throw attempts turn out to be the trip-ending ones. Pairs of free throws, three-shot foul trips, the occasional technical foul and and-ones all mix together, and the average works out to roughly 0.44 of an attempt per possession ended. Analysts therefore multiply free-throw attempts by 0.44 rather than counting them one for one. The factor is an estimate that fits the way the modern game is played; it is not a law of nature, and slightly different versions (0.40, 0.45) have been used in different eras and by different data providers.
Since the arrival of full play-by-play data and tracking cameras, the guessing has become less necessary. Modern data providers can mark the exact event where one possession ends and the next begins. But the estimate remains useful for two reasons. It still works for any season or league where only box-score data exist, which includes most of basketball history. And it is accurate enough that the direct count and the estimate usually agree to within a possession or two a game. If you want to understand where the numbers on this site come from, the data sources page explains what is recorded directly and what is calculated.
The arithmetic of a single shot
The reason efficiency beats volume is plain once you put numbers on individual shots. Take the three most common ways a possession can end well and compare them by points per attempt.
| Shot type | Make rate | Points per attempt |
|---|---|---|
| Long two-pointer | 40% | 0.80 |
| Average two-pointer | 50% | 1.00 |
| Three-pointer | 33% | 0.99 |
| Three-pointer | 36% | 1.08 |
| Shot at the rim | 65% | 1.30 |
| Pair of free throws | 75% each | 1.50 per trip |
These figures are illustrative, not a statement about any one season. They are chosen to show the shape of the problem. A long two-pointer that goes in four times out of ten is worth 0.8 points per shot. A three-pointer that goes in only one time in three is worth almost exactly one point per shot, and one that goes in a little more often is worth a good deal more. A shot at the rim that drops two times out of three is worth far more than either. Free throws, which get the shooter an uncontested attempt, are the most valuable possession-enders of all, because a typical pair is worth well over a point and a half.
This is the whole logic behind why teams stopped taking midrange shots. A possession is a fixed resource. Spending it on the lowest-value shot available is a poor use of the resource, even when the player taking the shot is good at it. A good midrange shooter who makes 45 percent of his long twos is producing 0.9 points per attempt. If that shot replaces a three-pointer from a teammate who makes 36 percent, the offense has given up nearly 0.2 points per trip. Over a season, thousands of such trips add up to a measurable gap in wins.
It also explains the three-point revolution in the numbers. The shot did not spread because it is glamorous or because fans like it. It spread because, at the make rates good shooters actually achieve, it produces more points per possession than most twos. Once coaches counted that way, the rest followed.
Why volume misleads
Return to the two 25-point scorers. Suppose the first scores his 25 points on 17 field-goal attempts and 7 free-throw attempts, with two turnovers. The second scores his 25 on 25 field-goal attempts and 3 free-throw attempts, with four turnovers.
Using the standard possession estimate, the first player used about 17 plus 3 (that is, 0.44 of 7) plus 2, or roughly 22 possessions. The second used about 25 plus 1 plus 4, or roughly 30. The first produced about 1.14 points per possession he used. The second produced about 0.83. On a team that otherwise averages around 1.1 points per possession, the first scorer is helping and the second is quietly hurting, even though the headline reads identically.
Volume misleads in three specific ways.
- It ignores the shots a player missed. A shooter who goes 8 for 25 has not simply scored 20 points. He has also used 17 failed attempts, and each failed attempt that ends in a defensive rebound is a trip that produced nothing.
- It ignores the ball a player gave away. Turnovers do not show up in a scoring average at all. They are the most direct subtraction from a team's possession count.
- It rewards whoever is allowed to shoot the most. A player on a team with a weak supporting cast is handed shots whether or not he converts them efficiently. A high scoring average can reflect opportunity, not skill.
None of this means that a high-volume scorer is a bad player. Some of the most valuable players in history were also among the highest-volume scorers, and the best of them kept their efficiency close to or above the league average while doing it. The point is that volume by itself is not evidence of quality. The useful question is always what the scorer gave back on the possessions he used. You can see the same idea in the shorthand statistics that compress it, such as true shooting percentage and effective field-goal percentage, which fold threes and free throws into a single shooting number.
From the player to the team
Efficiency matters at the team level for the same reasons, with one added wrinkle: a team has two ends of the floor. Its points per possession on offense and the opponent's points per possession against it are two halves of one scoreboard. Subtract one from the other, per 100 trips, and you get net rating, which is the cleanest single measure of how a team performs. You can read more about why in the article on net rating.
Pace belongs in the same discussion because it is what makes the possession count differ from game to game. A team that likes to run will generate more possessions for both teams. That raises the total points scored by both teams without telling you anything about who is better. A good habit when reading standings or leaderboards is to ask whether the number has been adjusted for the number of trips. The article on pace covers how tempo is measured and how it differs by era.
On the standings and leaders pages, you will often see per-game figures sitting beside rate figures. The per-game figures tell you what happened in a typical night. The rate figures tell you how well the work was done. When the two disagree, the rate figure is usually closer to the truth about quality, and the per-game figure is closer to the truth about workload.
Where points per possession gets complicated
A possession-based view is a strong tool, but it is not a perfect one, and it helps to know where it bends.
A possession can be more than one shot. When a team grabs an offensive rebound, the trip continues and may include several attempts. The point total for the possession is then the sum of everything that happened. This is one reason offensive rebounding is valuable: it gives the offense a second chance at the same trip. The trade-off with getting back on defense is discussed in the article on offensive rebounding versus getting back.
Not all possessions are equal. A possession that begins after a made basket is different from one that starts with a steal. Early in the shot clock, defenses are often unsettled, so fast-break trips produce high points per possession. Late in the clock, desperation shots drag the average down. A team that is forced into many late-clock possessions will look less efficient than one that gets to choose, even if its half-court execution is equally good. Analysts who split possessions by play type, such as transition, pick-and-roll or isolation, are trying to adjust for exactly this.
Context matters. Garbage time inflates and deflates numbers. Strategy at the end of a quarter, such as holding the ball for the last shot, changes the value of a trip. And a team trailing by a few points at the end of a game will intentionally foul, turning some possessions into free-throw trips. These distortions are small in a full season of data but can mislead in a single game.
Estimates carry error. The 0.44 factor is an average. In any given game it may be a little too high or too low. For seasonal totals the error mostly washes out, but a one-night possession count built from a box score can be off by a trip or two.
Efficiency is not the same as value. A player who takes only the easiest shots, such as layups created by others, can post a spectacular efficiency number while using very few possessions. That is useful, but it is different from a player who creates difficult shots for himself and still converts at an average rate. The usage side of the question is covered in the article on usage rate. Efficiency tells you how well possessions were spent; usage tells you how many were spent.
A coach's view of the same idea
It would be easy to think of points per possession as a product of the computer age. In fact, coaches have been working this way for a long time. Dean Smith, who coached North Carolina for 36 seasons, is widely remembered for tracking points per possession in his program rather than looking at raw scores. His reasoning was the same as everything above: a game can be won or lost with 70 points or 100, and the scoreboard tells you little about whether the offense did its job unless you know how many trips there were. His emphasis on taking good shots in his system, and on running the late-game spread offense that became known as the four corners, was an argument about maximizing what each trip was worth.
The modern analytics era made the argument public. Dean Oliver, an analyst who later worked inside NBA front offices, published Basketball on Paper in 2004, laying out how to count possessions and evaluate players and teams on a per-possession basis. His framework grew into the so-called four factors, which break a team's per-possession success into shooting, turnovers, rebounding and free throws. If points per possession is the idea, the four factors are the list of levers a team can pull to improve it.
Using the idea as a fan
You do not need a spreadsheet to put the possession view to work. A few habits are enough.
- When someone cites a scoring total, ask how many shots it took. A rough check is to compare points to field-goal attempts plus a little under half of the free-throw attempts. If a player scores more points than that total, he is above one point per possession used.
- When comparing teams from different eras, compare per-100 ratings or points relative to the league average rather than raw scoring. Scoring in the early 1960s was inflated by an extremely fast tempo; scoring in the late 1990s was depressed by a very slow one.
- When you watch a game, notice how a possession ends. A turnover, a contested long two, a shot at the rim, a drawn foul. Each ending carries a typical value, and the best offenses are those that reliably produce the high-value endings.
- When you hear that a player is an inefficient scorer, ask whether the claim is about shot quality, free-throw trips, turnovers, or all three. The problem may be a bad shot selection habit that is fixable, or it may be a reflection of the difficulty of the role.
What to take away
A basketball game is a fixed number of possessions shared between two teams. Scoring more than your opponent means getting more out of each one, and the cleanest way to see that is points per possession. Raw point totals reward tempo and opportunity; per-possession numbers reward the work. A midrange jumper, a three, a layup and a pair of free throws are not equal, and the difference between them is the reason teams changed how they shoot.
The next time a scoring average catches your eye, give it a second look. Ask what it cost, how many shots, how many turnovers, how many trips. The answer often changes which player looks like the star. For a quick sense of how the league's best offenses and defenses rank on this basis, the standings page is a good place to start.