Game Score is a formula published by John Hollinger that compresses one player's entire box-score line into a single number, scaled so it reads roughly like a points total. It is not an official NBA statistic. It does not appear in the league's own stat glossary, which lists hundreds of tracked and derived metrics without including it. It is a published analytical construct that became widely used because it is simple, transparent, and computable from any complete box score.
This calculator implements that published formula exactly, from the twelve box-score categories it uses, and adds the supporting figures a single line usually needs beside it. Arb Digital builds these tools because the value of a composite statistic depends entirely on knowing what went into it, and a Game Score quoted without its inputs is a number nobody can check.
What This Basketball Game Score Calculator Does
It takes a full single-game box-score line — shooting, rebounds split into offensive and defensive, assists, steals, blocks, turnovers and fouls — and returns the Game Score those figures produce. It also derives the points total from the shooting inputs rather than asking for it separately, which removes a common source of inconsistency where a hand-entered points figure does not match the makes it was supposedly built from.
The results grid adds three things the raw rating does not tell you. Game Score per 36 minutes rescales the figure to a common playing-time baseline, so a bench player's 22 minutes can be compared with a starter's 38. Total rebounds is shown so the split you entered can be checked against the box score's own total. True shooting attempts, defined as field goal attempts plus 0.44 times free throw attempts, is the standard published measure of shooting volume and is the denominator that the shooting-efficiency metrics use.
How to Use It
- Take the numbers from one game, for one player. Game Score is defined for a single performance. Summing several games and running the formula once is not the same as averaging the individual Game Scores, because the formula is linear but the per-minute rescaling is not.
- Enter field goals made and attempted including threes. Then enter three-pointers made separately. The three-point figure is used only to work out the points total, not as an extra shooting term.
- Split the rebounds. Offensive rebounds are weighted at 0.7 and defensive at 0.3, so a box score that only gives a combined total cannot produce an exact figure.
- Enter fouls and turnovers. Both carry negative weights. They are the terms people most often leave out, and leaving them out inflates the result.
- Add minutes for the per-36 figure. Minutes play no part in Game Score itself; they only scale the secondary number in the grid.
The Formula and How It Is Calculated
The published Game Score formula is a weighted sum:
GmSc = PTS + 0.4×FG − 0.7×FGA − 0.4×(FTA − FT) + 0.7×ORB + 0.3×DRB + STL + 0.7×AST + 0.7×BLK − 0.4×PF − TOV
Worked through with the values loaded on this page. Points first: 14 field goals of which 4 were threes gives 10 twos and 4 threes, so 20 + 12 + 8 free throws = 40 points. Then term by term: 40, plus 0.4 × 14 = 5.6, minus 0.7 × 25 = 17.5, minus 0.4 × (10 − 8) = 0.8, plus 0.7 × 3 = 2.1, plus 0.3 × 7 = 2.1, plus 2 steals, plus 0.7 × 5 = 3.5, plus 0.7 × 1 = 0.7, minus 0.4 × 2 = 0.8, minus 3 turnovers. Running total: 45.6, 28.1, 27.3, 29.4, 31.5, 33.5, 37.0, 37.7, 36.9, 33.9. Game Score is 33.9.
Over 38 minutes that is 33.9 × 36 ÷ 38 = 32.12 per 36 minutes, and true shooting attempts are 25 + 0.44 × 10 = 29.4.
Reading the Scale Without Over-Reading It
Hollinger scaled Game Score so that it lands in a range familiar from points totals: a figure around 10 corresponds roughly to an average starter's evening, 20 to a strong game, and 40 to something exceptional. That scaling is a presentational choice, not a statistical property, and it is the reason the metric caught on. It also creates the main way people misuse it, which is treating the number as though it were points.
It is not points. A player can post a high Game Score on a modest scoring night through rebounds, assists and steals with few turnovers, and a player can score 30 on 30 shots and finish with a mediocre one because the negative shooting term is heavy. The 0.7 penalty on every field goal attempt is the single most influential coefficient in the formula: each miss costs 0.7 and each make nets 0.4 plus the points, so volume without efficiency drains the total quickly.
What the Formula Deliberately Leaves Out
Game Score is a box-score metric, and it inherits every limitation of the box score. It contains no measure of defensive positioning, screening, spacing, or any of the off-ball work that does not generate a countable event. It cannot see who a player was guarding. Blocks and steals are the only defensive inputs, and both are known to be noisy indicators that reward gambling as much as sound defence.
It also makes no adjustment for context. A game played at a high pace produces more possessions and therefore more of everything, so two identical performances in different tempos score differently. It does not adjust for opponent, home court, or the score situation, which matters because garbage-time production counts the same as a fourth-quarter possession in a one-point game. And it is unweighted for era: a formula with fixed coefficients applied to a league where three-point volume has changed enormously does not treat 1995 and 2025 the same way, even though the arithmetic is identical.
Game Score, PER and the Shooting Metrics
Hollinger also published the Player Efficiency Rating, which is a season-level, pace-adjusted, league-normalised metric with a much more involved derivation. Game Score is often described as its single-game cousin, and the two share a family resemblance in which categories are rewarded, but they are not the same calculation and Game Score is not normalised to a league average of 15 the way PER is. Quoting one as though it were the other is a common error. The definitions of both, along with the 0.44 free-throw factor used here, are set out in the Basketball Reference statistics glossary.
The shooting-efficiency metrics sit alongside rather than inside it. Effective field goal percentage adjusts field goal percentage for the extra value of a three-pointer, and our effective field goal percentage calculator covers that definition. True shooting percentage goes further by including free throws in the denominator through the 0.44 factor this page also reports, and the true shooting percentage calculator handles it. The two are genuinely different measures — eFG% ignores free throws entirely and TS% does not — and neither is a component of Game Score, though both help explain a Game Score that looks surprising.
Comparing Players Fairly With a Single-Game Metric
Because Game Score is a raw counting-based total, minutes dominate any comparison made with it. A player who plays 40 minutes has roughly twice the opportunity of one who plays 20, and the metric does nothing to correct for that. The per-36 figure in the grid is the standard published rescaling, and it is the right basis for comparing players in different roles — with the caveat that per-minute rates from short stints are volatile, and extrapolating a 9-minute burst to 36 minutes produces a number nobody should defend.
The other honest limitation is sample size. One game is one game. A single Game Score describes an evening; it does not describe a player. If you are trying to rank performances across a season, the useful thing is the distribution — median, spread and how often a player clears a threshold — rather than any single figure. Our weighted average calculator and z-score calculator handle that kind of aggregation, and the Elo rating calculator covers a different approach to rating that updates on results rather than box scores.
Arb Digital builds free calculators like this one as search-visible entry points, then turns the traffic they earn into audience and enquiries with content and technical SEO built around the terms your readers actually type.
Content Marketing Talk to Arb DigitalCommon Mistakes to Avoid
- Entering twos and threes as separate field goal totals. Field goals made and attempted already include three-pointers. Adding them again double-counts the shooting terms.
- Using a combined rebound total. Offensive rebounds are weighted at 0.7 and defensive at 0.3. A single total cannot be split by the formula.
- Omitting fouls and turnovers. Both are negative terms, and leaving them blank produces a systematically inflated rating.
- Calling it an NBA statistic. It is Hollinger's published formula. The league's own glossary does not list it, and no official standings or awards use it.
- Comparing raw totals across very different minutes. Game Score is a counting metric. Without the per-36 rescaling, playing time drives the comparison.
Related Free Tools From Arb Digital
For the shooting side of the same box-score line, the effective field goal percentage calculator and the true shooting percentage calculator cover the two published efficiency definitions and how they differ. The winning percentage calculator handles team records, the Elo rating calculator covers results-based rating, and the z-score calculator and weighted average calculator are the general statistics tools for placing one performance inside a distribution. Browse the free online tools hub for the rest.
Frequently Asked Questions
It is a formula published by John Hollinger that compresses a single game's full box-score line into one number, weighted so the result reads on a scale familiar from points totals. It is an analytical construct, not an official league statistic.
No. It is Hollinger's published formula. The NBA's own stat glossary lists hundreds of tracked and derived metrics and does not include Game Score, and no official standing, award or record uses it.
Points, plus 0.4 times field goals made, minus 0.7 times field goals attempted, minus 0.4 times missed free throws, plus 0.7 times offensive rebounds, plus 0.3 times defensive rebounds, plus steals, plus 0.7 times assists, plus 0.7 times blocks, minus 0.4 times personal fouls, minus turnovers.
Hollinger scaled it so that roughly 10 reads as an average starter's night, around 20 as a strong game and around 40 as an exceptional one. Those are descriptive conventions from the metric's own scaling rather than defined thresholds.
No. PER is a season-level, pace-adjusted metric normalised to a league average of 15. Game Score is a single-game weighted sum with fixed coefficients and no normalisation. They share a family resemblance and are frequently confused.
Every attempt carries a 0.7 deduction while a make returns 0.4 plus the points scored. That structure makes the metric reward efficiency: a high-volume, low-percentage scoring night can produce a modest rating despite a large points total.
Not directly. It is a counting metric, so minutes dominate the comparison. The per-36-minute rescaling is the standard adjustment, though per-minute rates from very short stints are volatile and should not be extrapolated far.
Only through steals and blocks, which are the sole defensive events a box score records. Positioning, contest quality, screening and off-ball work are invisible to it, so it should not be read as a defensive rating.
This calculator reproduces a published basketball statistic for information and analysis only. Game Score is John Hollinger's formula rather than an official league statistic, it is not adjusted for pace, opponent or era, and it is not a measure of a player's overall value. Official statistics are those published by the league or competition concerned.
