All scores →
Loading…
Loading…
Loading…

Player efficiency rating: what it measures and what it misses

By 4Basketball Staff · · 12 min read

Player efficiency rating packs a player's whole box score into one per-minute number scaled to a league average of 15. Learn how it works, what it rewards, and where it misleads.

For a long time, comparing basketball players meant comparing lines of numbers. One player averaged 22 points, 8 rebounds and 3 assists; another averaged 18, 11 and 5. Which was better? The box score could tell you what each did, but not how to weigh the pieces against one another. Is a rebound worth more than an assist? Is a missed shot a small cost or a large one? Is a steal worth a block? Fans argued about it, but without a method the argument had no resolution.

Player efficiency rating, known as PER, was one of the first widely used attempts to resolve it. Created by the writer and analyst John Hollinger in the early 2000s, it takes the standard box-score categories, assigns each a value, adds up the good and subtracts the bad, adjusts for the pace of the team and the minutes played, and then scales the result so that the league average is always 15. A single number comes out. Because it was easy to compute from the box score, it spread quickly, appeared on major statistics sites and became the first advanced statistic that many fans learned.

PER is still quoted, and still argued over. This article explains what goes into it, how to read the scale, what it does well, and where its design leads it astray. It also places PER among its successors, which attempted to fix its shortcomings.

The idea behind an all-in-one number

An all-in-one statistic tries to answer the question of overall productivity. The reasoning is simple. Each box-score event contributes to winning or losing in some average way. A made field goal helps; a missed one hurts a little, because it uses a possession without scoring; a turnover hurts more; a rebound, an assist and a steal help. If you could assign a fair value to each, you could add them up.

The challenge is assigning the values. Hollinger derived his weights from the logic of possessions and the value of the average possession, and the details are complicated. In outline, the formula does the following.

  • Rewards made field goals (with extra credit for threes), made free throws, assists, offensive rebounds, defensive rebounds, steals and blocked shots.
  • Penalizes missed field goals, missed free throws, turnovers and personal fouls.
  • Adjusts the raw per-minute total for the pace of the player's team, so that a player on a fast team does not benefit simply from extra possessions.
  • Scales the result so that the league-wide average is 15.0 in every season, with all players' minutes taken into account.

The output is a per-minute rate, not a total. That is a deliberate choice. A player who plays 15 minutes a night and produces at a great rate can have a high PER, even though his contribution to the team's win total is limited by his playing time. Hollinger himself emphasized that PER measures per-minute production, and that a reader should look at minutes played alongside it.

How to read the scale

The fixed average gives PER an intuitive scale, at least as a rule of thumb. A PER of 15 is a league-average player. Starters tend to run above that. A season of 20 or more usually belongs to an All-Star-caliber player, and a season around 25 or higher is in the range of an MVP candidate. The very highest figures in league history, in the high 20s and above, belong to a small number of the greatest seasons ever played.

Those are rough bands, not official categories. The point is that the scale gives you something to compare against. If a player's PER is 12, he is producing a little below the average per minute. If it is 28, he is producing far above it.

A few features of the scale are worth noting.

It is normalized each year. Because the league average is set to 15 in every season, PER values can be compared across eras to a degree. A 25 in 1975 and a 25 in 2015 both mean "well above the average player of that year." This is a strength of the approach. It is also a limitation, as the average player of 1975 and 2015 are not the same, and the statistic cannot tell you which era's average was better.

It rewards sustained volume of good plays. Because it counts positive events and subtracts negative ones, a player who does many things well at a high rate scores highly. This is why big men who score, rebound and block shots at a high rate often have high PERs, and why players with a narrow role tend to have lower ones.

It sits at or near the top for the greatest scorers. Among players with long careers, Michael Jordan sits at or near the top of the career list. His career is the subject of the article on Michael Jordan. Wilt Chamberlain's seasons, covered in the article on Wilt Chamberlain, also stand out when the statistic is computed for his era, though some categories such as blocks and steals were not officially recorded until the 1973-74 season, so the formula has incomplete inputs for earlier years.

What PER does well

PER earned its popularity for good reasons.

It is accessible. It requires only the box score, which means it can be computed for any season in which the standard categories were recorded. Casual fans can look at one number and form a rough view of a player's production.

It adjusts for context better than raw totals. The pace adjustment means that a player on a fast team does not look better than an equivalent player on a slow team. The per-minute basis means that a bench player's production can be compared with a starter's. Both are real improvements over per-game averages. The reason pace matters is explained in the article on pace.

It captures a wide range of contributions. A scorer who also rebounds and passes will score higher than a pure scorer of the same volume. The all-in-one framing encourages thinking about the whole game.

It correlates with how good players are generally considered to be. Players with the highest PERs in a season are typically those who are in the MVP conversation. When a statistic agrees with the consensus in most cases, it is at least capturing something real. The disagreements are where the interesting questions lie.

What PER misses

The weaknesses of PER are as well known as its strengths, and a careful reader should understand them.

It undervalues defense

The central criticism is that PER is built almost entirely from offensive and rebounding events, with only steals and blocks as direct measures of defense. But much of defense does not appear in the box score. Positioning, communication, contesting shots without blocking them, forcing a drive to the wrong side, preventing a pass from being thrown at all: none of these generates a statistic. A player who is an outstanding defender but not a high-volume shot-blocker or thief will have his defensive contribution largely invisible to PER. Hollinger acknowledged the limitation, and it is part of the broader problem described in the article on defensive metrics.

The effect is a systematic bias. Players whose value is mostly defensive, such as perimeter stoppers and defensive anchors who do not accumulate blocks, tend to rank lower on PER than their influence on winning would suggest. The reverse is also true: players who pile up steals and blocks may have their defensive value overstated, since gambling for steals can hurt a defense even while it inflates the statistic.

It rewards shooting volume

The way PER treats missed shots matters. A missed field goal is penalized, but not as heavily as a made field goal is rewarded. The weights are set so that a shot attempt at around the league's average efficiency is roughly neutral. Hollinger's own reasoning was that the penalty should reflect the lost possession, but critics have argued that the balance of the formula is generous to high-volume shooters, who can accumulate positive value by taking many shots at a moderate efficiency. Because a shooter of average efficiency contributes more than the formula appears to charge him for the possessions he consumed, a player can raise his PER by shooting more, even if each additional shot is of mediocre quality.

That makes PER somewhat sensitive to usage. A player who is given a larger role can see his PER rise even if he becomes no more effective per possession. The relationship between usage and efficiency is covered in the article on usage rate. In practice, the tendency is modest compared with the other effects, but it is why analysts have been cautious about comparing PER among players with very different usage.

It ignores what is not in the box score

A great deal of offense does not appear as a statistic. Setting screens, spacing the floor, drawing a defender away from a cutter, making the right read on a pass that is not an assist: all of this changes the success of a possession and none of it shows up. Players whose value comes from these things, including good screeners and off-ball movers, are undercounted. The influence of such players is the sort of thing that on/off and lineup measures attempt to capture, with their own problems, as described in the article on plus-minus, on/off and why the raw number lies.

It treats events as if every one has the same value in every context

A rebound is a rebound in the formula. But a rebound that comes in traffic against a crowd is not the same as one that falls to a player alone under the basket, and a teammate may have boxed out the opponent to let him have it. In the same way, a made basket in the first quarter of a blowout counts the same as a made basket in the final minute of a tie game. This is a general limit of any box-score statistic. The article on clutch statistics takes up the context problem for late-game situations.

It does not separate the player from his teammates

PER is computed from a player's own box-score numbers, but those numbers are affected by who else is on the floor. A player on a team with excellent passers will get more open shots. A rebounder on a team with a poor rebounding supporting cast will get more opportunities. The statistic cannot untangle these effects.

Why the weights are debatable

Underneath all of these criticisms is a deeper question: how should the weights be set? There is no single correct answer. Hollinger's choices represented one reasonable set of judgments about the value of events. Other analysts, using regression methods that relate box-score events to team results, arrived at different weights. For example, later measures give less credit to scoring volume and more credit to efficiency and to defensive contributions, and they treat rebounds and assists differently.

This is not a flaw peculiar to PER. Any all-in-one statistic has to weight events, and the weights embody a theory of what matters. The more transparent the theory, the easier it is to criticize, which is perhaps why PER has attracted so much criticism: its construction is relatively open. A reader who understands the weights can see which kinds of players are being favored and adjust.

The statistics that followed

Since PER appeared, several families of all-in-one measures have been developed, many of them designed to answer its shortcomings.

  • Win shares, built on a framework developed by the baseball analyst Bill James and adapted to basketball, allocate a team's wins to individual players using their offensive and defensive contributions. They give more explicit credit for defense, though still derived from box-score data.
  • Box plus-minus (BPM) uses a regression of box-score numbers against a more sophisticated measure of impact to estimate a player's contribution per 100 possessions relative to an average player.
  • Value over replacement player (VORP) turns BPM into a cumulative total by comparing a player with a replacement-level one and scaling by minutes played.

These measures are explained in the article on win shares, BPM and VORP. They are not free from criticism either, but they reflect a decade of refinement. Alongside them are metrics that incorporate on/off or tracking data to try to capture defense more directly. For a plain-language tour of the whole alphabet, the basketball stats glossary lists them with short explanations.

Using PER sensibly

Despite its flaws, PER can still be useful if it is used with an awareness of what it is.

  1. Treat it as a measure of box-score production per minute. That is what it is. It is not a measure of total value, and it is not a measure of defense.
  2. Compare it with minutes played. A high PER over a small number of minutes is less reliable than a high PER over a full season of starter minutes. For small samples, remember the lessons of the article on small samples.
  3. Look at shooting efficiency alongside it. A quick check of true shooting percentage can reveal whether a high PER rests on efficient scoring or on volume.
  4. Be suspicious of large PER gaps between defensive specialists and offensive stars. The gap partly reflects what the formula can and cannot see.
  5. Use it to start a conversation, not to end one. If a player's PER disagrees with your eye, that is worth investigating. Sometimes the statistic is revealing something you missed; sometimes your eye is seeing something the statistic cannot.
  6. Check the context of the team. A player on a team that plays fast, shoots a lot and has few other options can post a higher PER than an equivalent player on a deliberate team with several scoring options.

PER is also easy to compare across seasons for the same player, which makes it useful for tracking development and decline. A young player whose PER rises from 13 to 17 to 20 is clearly growing into a larger role. An older player whose PER falls steadily is probably losing some of his production. The study of such trajectories is covered in the article on aging curves.

A fair verdict

It is easy to be dismissive of PER now that better measures exist, but that would be unfair. When it was introduced, it was a significant step forward. It encouraged fans to think about efficiency and the whole box score instead of points per game. It forced a conversation about how to weigh different contributions. And it set a template for later measures: take the available data, assign defensible weights, adjust for context and produce a number that can be compared.

Its flaws are mostly the flaws of its inputs. The box score does not record defense well, and no formula built on it can fix that. The honest conclusion is that PER is a useful summary of offensive and rebounding production, a weak guide to defense, and a decent first filter for identifying players who deserve a closer look. The leaders and players pages on this site show many of the underlying statistics, so you can form a view of your own.

What to take away

PER takes everything in the box score, weights it, adjusts it for pace and minutes, and scales it so the league average is 15. It is quick, intuitive and broadly consistent with the way experts rank players. It also leans on offense, underrates defense that does not generate steals and blocks, and can reward volume over quality when shots are plentiful.

Use it the way you would use any single summary of a complicated thing: as a starting point. A high PER tells you that a player produces a lot on the court per minute. It does not tell you whether he makes his team better in the ways that matter most, and for that you need to look at efficiency, role, defense and the team around him.

Search more on this

Keep reading

Pace: possessions per game and why it changes everything

Pace counts how many possessions a team plays per game. Learn how it is measured, how tempo has swung across eras, and why ignoring it distorts nearly every counting statistic in basketball.

4Basketball Staff · Jul 19, 2026 · 12 min read

Expected points and shot quality: tracking-era basketball

Tracking cameras let analysts estimate how likely each shot was to go in. This guide explains expected points, shot quality versus shot making, and the limits of the models behind the numbers.

4Basketball Staff · Jun 27, 2026 · 11 min read