How many times should you watch a player before writing a recommendation?

More than once, and on different occasions: different opposition, different scorelines, ideally a match the team loses. A single viewing largely measures form on the day and the level of the opposition. What matters is not the number itself but the spread between viewings.

Should you trust the data or the eye?

Both, and preferably not separately. Subjective coach assessment has predictive value over and above objective testing, and the combination beats either alone. What does not work is letting the numbers decide without context, or letting instinct write the report and then fetching numbers that confirm it.

How do you assess a player who is small for their age?

Assess against biological peers where you can, not only against chronological ones. Later-maturing players are deselected at scale, but those who survive selection anyway show better learning strategies — they have had to solve situations with something other than their body. Write that in the report rather than compensating in the rating.

For scouts

Two scouts agree on the present and disagree on the potential

8 min read

When 83 scouts were asked to rank the same 24 young players they broadly agreed on who was best right now — and almost entirely disagreed on who would become best. That is not an argument for giving up on potential. It is an argument for writing it down separately, with lower confidence, so that it can be reviewed later.

Two fields, different confidence

The most important change to a report template is also the cheapest: split current level and potential into two fields, and explicitly mark the second as a guess. Scouts describe their work as structured but in practice reach the decision through an intuitive weighing — and under time pressure fall back on whatever is quickest to see, which is physique and pace.

It is also worth knowing how weak the prediction is in itself. In the largest longitudinal dataset in the field, tactical skills, kicking technique and sprinting are the strongest individual predictors — but even the best model explains only part of the outcome. And the overlap between successful juniors and successful seniors sits at around seven per cent.

Put the birth month at the top

Two biases bite harder than all the others, and they reinforce each other the further up the selection chain you go: relative age and biological maturity. The maturity bias is the stronger of the two — the difference between two players born the same year can be several years in development, while birth month gives at most one year.

At Swedish regional camps around 72 per cent of the boys selected were born in the first half of the year, without the coaches themselves seeing the skew. They confused physical maturity with talent. Writing the birth month, and where possible maturity status, at the top of the report costs one line and means whoever reads the rating reads it against the right background.

Anchoring: the first player you watched sets the scale for everyone else that day. Vary the order deliberately, and never judge from the first clip you see of a player.
Geography: players from smaller towns need to perform noticeably better than city players to be selected. That is measured, not a feeling.
Position: the relative age effect bites differently by position, and is strongest among centre-backs.
Halo: a strong impression in one aspect colours every other judgement. Rate one attribute at a time instead of forming an overall impression.

Decide what you are looking for before you press play

Introducing a structured form is not enough on its own — a Dutch study found that structured data collection with mechanical aggregation produced no higher reliability than unstructured judgement. What makes the difference is what you decide in advance, not how the form looks.

So write down which two or three attributes this assessment is about before you start watching, and judge them one at a time. When recruiters rank what they actually look for in thirteen-year-olds, technique under pressure, decision-making and coachability come top, while physique and anthropometry are deliberately set aside. That is easy to say and hard to do once the match is running.

What video can and cannot do

Scouts often believe an in-person assessment is more accurate than one made on video. The research does not support that: the difference lies in time and workload, not in accuracy. What does differ is what the material shows.

A highlight reel shows finishes and skills, that is, moments where the player already has the initiative. It says almost nothing about positioning, scanning or what the player does in the sixty seconds before. So ask for something more, and say what you want — otherwise whoever films will guess wrong.

Clips that start five to ten seconds before the touch, so the build-up, the scanning and the positioning are visible.
A wide enough frame that the space is visible. A clip zoomed in on the player cannot be used to judge positioning.
At least one full half alongside the reel, ideally against strong opposition.
At least one clip where something goes wrong, with the camera left running for ten more seconds.
The context written out: level of opposition, position, minute, scoreline and date.

Write the report so it can be read in two years

Write descriptively and from examples: what happened, in which minute, against whom. Avoid labels — wonderkid, the next somebody, winner’s mentality. A label feels like information when it is written and is useless when it is read, because nobody can say what would disprove it.

Use a small recommendation scale rather than a fine-grained one. Monitor, watch again, invite, proceed. Four steps force a decision; a ten-point scale lets you write a 6 and avoid taking a view.

Have two people assess independently and compare only afterwards. Date every report, and once a year do the thing nobody has time for: read your own old reports against what actually happened. It is the only feedback loop this job has, and it costs nothing.

What applies when the player is a minor

Assessment is one thing, contact is another. In Swedish football it counts as an illegal approach to contact a player at another club directly — even when it is only an invitation to a trial, and even when the message is friendly. The order is fixed and always begins with the player.

In Breakout you can search on position, age, league and on what happens in the clips, and rate in five parts for players who have ratings switched on. Two things are worth knowing in advance: players under 13 never appear in search or the radar, and players between 13 and 17 are private and invisible to scouts until they switch visibility on themselves. If you want to make contact you send a request the player accepts or declines, and you never receive an email address or phone number.

Common follow-up questions

How many times should you watch a player before writing a recommendation?: More than once, and on different occasions: different opposition, different scorelines, ideally a match the team loses. A single viewing largely measures form on the day and the level of the opposition. What matters is not the number itself but the spread between viewings.
Should you trust the data or the eye?: Both, and preferably not separately. Subjective coach assessment has predictive value over and above objective testing, and the combination beats either alone. What does not work is letting the numbers decide without context, or letting instinct write the report and then fetching numbers that confirm it.
How do you assess a player who is small for their age?: Assess against biological peers where you can, not only against chronological ones. Later-maturing players are deselected at scale, but those who survive selection anyway show better learning strategies — they have had to solve situations with something other than their body. Write that in the report rather than compensating in the rating.