Goals per 90 isn't pure finishing
Goals per ninety minutes is supposed to be the fair way to compare forwards. Raw goal totals reward whoever plays the most, so dividing goals by minutes and scaling to a full match seems like the obvious correction — a rate that lets a rotation striker stand beside an ever-present one on equal terms. StatLine’s Premier League player tables now carry goals and minutes for every player, so the per-ninety rate is a simple sort away, and it is a genuinely useful number. But it gets quoted as a pure measure of finishing ability, a readout of how lethal a striker is, and it is not that. Goals per ninety folds penalties, sample size, finishing variance, and the quality of a player’s service into a single figure, and only a fraction of it belongs to the striker’s finishing at all.
What goals per 90 actually records
The calculation is exactly what it sounds like: total goals divided by minutes played, multiplied by ninety. A striker with fourteen goals in 1,800 minutes is scoring 0.70 per ninety. The rate strips out the advantage of simply playing more, which is a real improvement over the raw total, and it puts a supersub and a starter on a comparable scale.
But a goal in this calculation is a goal regardless of how it arrived — a tap-in from two yards, a penalty, a deflected shot that wrong-foots the keeper, and a thirty-yard screamer all add exactly one to the numerator. The rate treats them identically and asks no questions about difficulty, service, or luck. It measures how often the ball ended up in the net while a player was on the pitch, which is related to finishing but is not the same thing.
Penalties sit in the same column
The single cleanest way to inflate a goals-per-ninety figure is to take penalties. A designated penalty taker on a team that wins a lot of spot-kicks can add six or eight goals a season from twelve yards, and every one of them lands in the same column as goals scored from open play. Two strikers with identical open-play output can post very different per-ninety rates purely because one of them is trusted with penalties and the other is not.
Penalties are a real skill and a real contribution, but they are a different skill from finishing chances in the run of play, and they belong to a role — being the nominated taker — as much as to an ability. A goals-per-ninety ranking that does not separate them is partly a ranking of who stands over the penalty spot, which is a decision the manager made, not a measure of how well a striker finishes.
The denominator lies in small samples
Per-ninety rates are seductive precisely because they let low-minute players onto the leaderboard, and that is also their biggest trap. A substitute who scores twice in 200 minutes posts a gaudy 0.90 per ninety, a figure that looks elite next to a starter grinding through 2,500 minutes at 0.60 — but the substitute’s rate is built on a sample so small that a single goal swings it wildly.
Rate stats need a minutes threshold to mean anything, and goals per ninety without one is a machine for producing misleading leaders. The eye-catching per-ninety figures at the top of an early-season or low-minute list are mostly the product of variance finding a small denominator. The starter’s more modest rate, earned over a full season of matches, is the one carrying real information.
Finishing variance and xG overperformance
Even over a decent sample, goals per ninety is noisier than it looks, because finishing is a high-variance skill and goals are rare events. A striker can outscore his expected goals for a stretch through a hot run that regresses, or fall short of it through a cold one that will correct, and the per-ninety rate captures the streak as if it were the true level.
This is why a player’s goals per ninety can look elite for half a season and then sag — not because his finishing collapsed, but because the initial rate was partly variance. Expected goals per ninety is the more stable underlying signal, and comparing it to actual goals per ninety tells you whether a striker is genuinely at that level or riding a finishing streak the schedule has not yet corrected.
Service decides the ceiling
A striker cannot score chances he is not given, and the volume and quality of those chances depend far more on his teammates and system than on him. A forward in a dominant side that manufactures a stream of high-quality chances will post a goals-per-ninety rate a striker in a struggling team cannot approach, even if the second striker is the better finisher of the two.
Goals per ninety therefore encodes the strength of a player’s supply line as much as his own quality. Move an elite finisher to a team that creates little and his rate falls; drop an average finisher into a chance-creating machine and his rate climbs. The number moves with the service, and the service belongs to the eleven, not the one at the end of it.
What to read instead
Read goals per ninety beside the context that reveals where the goals came from. Non-penalty goals per ninety strips out the spot-kicks and isolates open-play finishing. Expected goals per ninety shows the quality of chances a striker is getting and whether his output is sustainable or a hot streak. A minutes filter — a full season’s worth before trusting a rate — keeps small-sample substitutes from topping the list.
StatLine’s Premier League tables make the first step easy: goals and minutes are there for every player, so the per-ninety rate is a quick calculation, and reading it against appearances and shot volume shows whether a striker is scoring at a rate his workload and chances actually support. The rate is the starting question, not the answer.
The honest read
Goals per ninety is a real and useful rate, and a striker sustaining a high figure over a full season on non-penalty goals is genuinely lethal — the number is not lying at the top of a large, penalty-adjusted sample. It is the fairest simple way to compare forwards who play different amounts, and it deserves its place in any scoring conversation.
But it is not pure finishing. It carries penalties, small-sample noise, finishing variance, and the quality of a player’s service inside it, and any one of those can move the rate as much as the striker’s own ability. Read goals per ninety with a minutes threshold, next to the non-penalty and expected-goals versions, and treat the raw rate as a headline that still needs the context underneath it before it means what people think it means.