Player Efficiency Rating is a scoring stat in disguise
Player Efficiency Rating arrived with a promise that was hard to resist: take everything a player does in a box score, weigh each act by its value, adjust for pace and minutes, and hand back a single per-minute number where 15.0 is always league average and 30 is an MVP. For two decades it has done the work a tired argument needs done — one figure to rank players across positions and eras, printed next to names as if it settled the question of who is better. But a single number that claims to capture all of basketball has to decide what basketball is, and PER decided, quietly, that basketball is mostly scoring. Read its leaderboard and you are not looking at a ranking of the most valuable players. You are looking at a ranking of high-volume, efficient scorers, with everyone else sorted underneath them.
What PER is actually built from
PER is a weighted sum of box-score events — points, field goals, free throws, rebounds, assists, steals, blocks, turnovers — converted to a per-minute rate and then normalized so the league average lands at 15.0 every season. The normalization is the part people remember and the weighting is the part that decides everything. Because the formula is anchored to scoring and the efficiency of that scoring, the events that move a player’s PER the most are the ones that fill the points column. A player who scores a lot on reasonable efficiency climbs almost automatically; a player who does the quieter work of a good possession barely registers. The number wears the costume of a complete summary, but its skeleton is a scoring rate with garnish.
The defense it cannot see
The deepest hole in PER is the same hole in the box score it is built from: defense barely exists in it. The only defensive events the formula can reward are steals and blocks, and those are a thin, misleading sliver of defending — the gambles and the swats, not the positioning, the rotations, the ball-pressure, or the shots deterred without a fingerprint. A wing who erases the opponent’s best scorer for forty-eight minutes, forces a dozen bad shots, and never gambles for a steal contributes almost nothing to his own PER on that end. A center who anchors an elite defense by walling off the rim without blocking much gets no credit for it. PER can tell you who scored efficiently. It is close to blind on who stopped anyone from scoring.
Why it rewards taking shots
PER’s weighting has a subtler bias that its creator openly acknowledged: it tends to reward shot volume as long as the shooter is roughly average or better at converting. A made basket is worth more than the cost the formula charges for the misses that come with chasing volume, so a player can lift his rating simply by taking and making a lot of shots at ordinary efficiency. That is why high-usage scorers populate the top of the leaderboard and why role players who do everything well except shoot a lot get sorted beneath them. The stat does not punish a chucker hard enough and does not reward a low-usage connector enough, which means it reads usage as value far more than it should.
The things it underweights
Some of the most valuable acts in basketball move PER hardly at all. The pass that comes one before the assist, the screen that springs a shooter, the box-out that lets a teammate grab the rebound, the cut that bends a defense, the closeout that turns a good look into a bad one — none of these are box-score events, so none of them touch a player’s rating. Even assists, which PER does count, are credited only to the final passer and weighted modestly, so a brilliant hub who creates the entire shot diet of a team is undersold relative to the teammate who finishes the plays. A metric that sees only the events with a tally mark will always undervalue the players whose best work never earns one.
It is worth being concrete about how lopsided the weighting gets. Because the formula is anchored to scoring, a player who adds a few points to his per-minute output climbs noticeably, while a player who improves his defense from poor to excellent — a swing that can decide playoff series — moves his PER almost not at all, since the box score gives the formula nothing to grab onto. The predictable result is that PER and the plus-minus impact models often disagree most about exactly the players who matter most: elite defenders, low-usage connectors, and stars whose value is spread across the parts of the game the box score never writes down. When two honest attempts to measure value diverge that sharply on the same player, the one built only from counting stats is the one to distrust.
What the single number hides
The appeal of PER is the appeal of any one-number summary: it ends the argument. But ending the argument is exactly the problem, because the act of compressing a player into a single figure forces choices about what matters, and those choices are buried where no one reading the leaderboard can see them. Two players with the same PER can be wildly different basketball players — one a scoring guard who plays no defense, the other a two-way forward whose offense is quieter — and the number erases the difference rather than explaining it. A summary stat is only as honest as its willingness to admit what it left out, and PER is most often quoted with no such admission, as though 22.4 were a fact rather than a recipe.
What to read instead
The honest move is to stop asking one number to do all the work and to read a small panel instead. For scoring, true shooting and usage together tell you how much a player scored and how efficiently — which is most of what PER was actually measuring anyway, stated plainly. For the all-around case, the plus-minus family of estimators folds in defense and on-court impact in a way the box score cannot, and on/off splits show whether the team is genuinely better with a player on the floor. None of these is perfect, and the plus-minus models carry their own noise, but read as a group they stop you from mistaking a scoring rate for a verdict. Keep PER as a quick, rough scoring proxy if you like, and never let it referee a debate about the best player.
The honest read
Player Efficiency Rating is a real and clever piece of work, and for the narrow job of flagging efficient high-volume scorers it does fine. The trouble is the job it is asked to do in public, which is to rank players outright, and at that job it is structurally compromised: it is dominated by scoring, nearly blind to defense, generous to shot volume, and stingy with the connective work that wins games. Treat it as a scoring stat in a value stat’s clothing, read it next to the impact metrics it cannot replicate, and remember that the most valuable player on a floor is sometimes the one PER has filed somewhere in the middle.