How to actually read an NBA box score
The box score is the oldest stats document in basketball, and it's survived a hundred different scoring philosophies, three significant rule changes, and the entire analytics revolution. It is also, line by line, one of the most misleading summary documents in professional sports — not because the numbers are wrong, but because the way most people read them is wrong. They scan top to bottom looking for the biggest number, decide who "had a good game," and stop. That isn't reading a box score. It's reading a leaderboard.
Every column has a story. A few of those stories are the ones the column is ostensibly telling. Most of them aren't. Here's a column-by- column field guide for what each number is really saying, with a small amount of opinion about which ones to trust and which ones to ignore.
Points (PTS)
The headline number, and the one that misleads most often. PTS is a product of three different things — how often a player shot, how accurately they shot, and how many free throws they earned — but the column collapses all three into one digit. Twenty-four points on twenty-six shots is a different game than twenty-four on thirteen. The column doesn't tell you which one happened. Always read points next to field-goal attempts, not alone.
Field goals (FG, FGA, FG%)
The most informative trio in the entire box score, and the one that usually gets glanced past on the way to PTS. The made-attempted pair is what tells you whether the points came efficiently or because somebody fired thirty shots. FG% is the cleaner read: above .500 is a very good shooting night for a guard or wing, anything over .600 is excellent, and a center under .500 had a quiet game even if the points look fine.
The catch: FG% counts a contested two-pointer the same as a wide-open layup, and it counts a three-pointer the same as a two. Modern analysts almost never look at raw FG% in isolation anymore for that reason. Which brings us to:
Three-pointers (3P, 3PA, 3P%)
The single column that has changed the most in the last fifteen years. Average three-point volume per team has roughly tripled since 2010, and 3P% is now arguably more predictive of long-term value than total FG%. League average from deep sits in the high thirties; anyone over .400 on real volume is genuinely excellent, anyone in the low thirties is hurting their team's spacing whether or not the points show it.
A useful trick: ignore percentage when attempts are below three. Small samples within a single game produce wild numbers — a player who goes 2-of-2 from deep is not actually shooting 100% in any meaningful sense.
Free throws (FT, FTA, FT%)
The most underrated number in the box score. Free-throw rate (FTA per field-goal attempt) is one of the cleanest signals in the game. It tells you whether a player is actually attacking the rim or just shooting jumpers. Scoring twenty-five points on five free throws is a very different evening than scoring twenty-five on twelve. The first guy floated; the second one was fouled hard repeatedly because he kept getting downhill.
FT% itself is more of a personality trait than a game-night stat. Most players are within a few points of their career number every night. It's worth checking the season number; rarely worth dwelling on the per-game value.
Rebounds (OREB, DREB, REB)
The total rebound column flatters big men who play a lot of minutes and is bad at distinguishing effort from opportunity. The split into offensive and defensive is much more informative. Offensive rebounds are scarcer and harder to manufacture, so a stat line of 14 boards with 6 of them offensive is a much louder game than 14 with 1 of them offensive. The latter is mostly the result of being seven feet tall in the right place.
Assists (AST)
Assists are scored by humans, in real time, with judgment, and the scorers are not consistent across arenas. Take the column with a grain of salt as a precise number; trust it as a rough indicator. What matters more than the count is the assist-to-turnover ratio, which you can read by combining this column with the next one.
Turnovers (TOV)
Turnovers are the most underweighted negative in the game. A turnover usually leads to an opponent transition possession that is more efficient than the average half-court possession. So a turnover isn't worth zero points; it's worth roughly one and a quarter, on average, in expected value. A point guard with 8 assists and 6 turnovers looks productive on paper and was probably net negative on the possession ledger.
Steals and blocks (STL, BLK)
Both columns are noisy. Steals reward gambling that worked; the same gamble that failed shows up nowhere on this row. Blocks are flashy but only count when the ball doesn't go in, which means a defender who alters a shot without touching it gets nothing. They're real skills, but the box score versions of them are a fraction of the full defensive picture.
Plus/minus (+/-)
The most contextual number on the page. A bench player whose +14 came in garbage time against the third unit had a different night than a starter whose +14 came against the opposing starters in close minutes. Single-game plus-minus is almost meaningless. Aggregated plus-minus across many games starts to mean something. Aggregated plus-minus with opponent and lineup adjustments — what's usually called RAPM or on/off — means a great deal.
The ten-second read
Suppose you have ten seconds and you can only look at three things. Look at FG attempts, FT attempts, and turnovers. Those three numbers together tell you how often a player had the ball, how aggressively they used it, and what they cost the team for the privilege. Points will follow. Almost everything else is downstream of those three decisions.
The box score isn't dishonest, but it's a summary, and like all summaries it leaves things out. The bigger your library of context for what's missing, the more accurately you'll read what's there. After a couple seasons of looking, the numbers start telling you the narrative — not the other way around.