Per 90 vs per 100 touches: which number should you trust?
Raw totals are almost useless in football data. A player with 40 completed dribbles might be a bench winger having a great season or a starter having a quiet one. You can't know without asking: 40 out of how much football?
There are two honest ways to answer that, and they measure genuinely different things. Knowing when to use which is one of the highest-value habits in data scouting.
Per 90: normalising by time
Divide any counting stat by minutes played, multiply by 90, and you get the stat's rate over a full match. This is the standard currency of football data, and for good reason: it lets a 1,200-minute squad player and a 3,400-minute ever-present be compared fairly.
What it answers: how much of this does the player produce per match played?
What it hides: involvement. Per 90 treats every minute as equal, but a winger at a possession-dominant club might see three times as much of the ball as one at a counter-attacking club. Their per-90 numbers are shaped by their teams as much as by themselves.
One caution: per 90 gets noisy at low minutes. A stat built on 400 minutes is a rumour, not a measurement, which is why FDscout applies minimum-minutes floors before ranking anyone.
Per 100 touches: normalising by involvement
Divide the stat by the player's touches instead, and scale to 100. Now team context largely falls away: you're measuring what a player chooses to do with the ball when he actually has it.
What it answers: style. Is this a player who dribbles, passes, shoots, crosses when given possession, regardless of how often possession arrives?
What it hides: volume. A player can have a striking per-100-touches profile on very little of the ball, and per-100 alone won't tell you his total contribution is small.
The same player, two different pictures
Take two leaders from our live 1v1 wingers ranking: Lamine Yamal and Henry Mosquera attempt a similar number of take-ons per 90 (8.95 vs 7.80). On per 90, near twins. On per 100 touches they separate completely: Mosquera attempts a take-on every six touches, Yamal every nine. Barcelona's possession game feeds Yamal roughly 2,000 touches across his minutes; Bragantino gave Mosquera 735. One player's appetite is spread across constant involvement; the other compresses the same aggression into scraps.
Neither number is "right". Together they tell you something neither could alone: what happens to each player's output if you move him into a different team's context. That's the actual scouting question.
Which one should you trust?
- Comparing output fairly across playing time: per 90. Goals, chances created, tackles: per 90 is the default lens.
- Diagnosing style and decision-making: per 100 touches. Dribble attempts, pass frequency, fouls drawn: per-100 shows the player's instincts with team context stripped out.
- Projecting a transfer: both. Per 90 tells you what he produced; per 100 touches tells you what he'd likely do with more (or less) of the ball.
How FDscout handles it
Every profile carries a Totals / Per 90 toggle, and the involvement-based stats (dribbles, passes, dispossessions, fouls drawn) are also computed per 100 touches. Everything is then ranked as a percentile against players in the same position across 30 leagues, so you never have to guess whether 7 take-ons per 90 is a lot (it is: roughly the 99th percentile for wingers).
The full vocabulary, covering every stat, both normalisations and how the percentile pools are built, lives in the stats glossary. And if you want to see the difference in the wild, open any two wingers from different league contexts in the comparison tool and flip the toggle.