What a cognitive profile measures, and what it does not

Technical talent is measured and tactical skill is measured. How a player decides, in the half second that matters, has been left to intuition for as long as the game has existed.

Where the numbers come from

Nobody sits the player down. The profile is derived from what he already did on the floor, across the possessions the database holds for his league and season.

That is the important distinction against every lab based approach: the measurement happens under real pressure, against real defenders, at the real speed of the competition.

It also means the profile arrives without asking anything of the club, and it updates as the season goes rather than freezing on the day of a test.

The four domains

Control is the ability to govern a possession: choosing the efficient solution, holding off the cognitive error, and keeping decision quality stable as the load rises.

Attention is perceiving and updating what matters while the possession is still running, and Prediction is acting on what is about to happen before it does, which at this speed is a structural advantage.

Resilience is the one staffs ask about most: holding output when the physical and cognitive pressure peaks, which is the part of a season that decides the other part.

A percentile, not a grade

A score of 75 means the player is above 75% of his league in that domain, in that season, which makes it a position in a field rather than a mark out of a hundred.

Prediction is normalised by role, using percentages instead of volumes, so a point guard and a centre can be compared without physicality doing the work.

The scale is read in bands: the top fifth of a league is a structural strength, the middle is a development area, and the bottom is a priority rather than a verdict.

Team DNA, and why fit is not just talent

The individual profiles combine into the cognitive signature of a squad, which is the way a team systematically thinks and plays rather than the way it is described in a preview.

A signing then gets read against that signature twice: similarity says whether he thinks the way the system needs, affinity says whether he covers the domains the roster is thin in.

Being good is not enough if neither is true, which is the polite way of describing several expensive summers.

Where it stops

A cognitive profile is not a personality assessment, and it does not predict how someone will behave in a locker room or react to a coach.

It also depends on having enough possessions: the profile covers more than 39,500 players rather than the whole database, because a thin sample would produce a confident number with nothing behind it.

Read as one axis among several it is the most useful new thing on the table. Read as a verdict on a person it becomes exactly the kind of overreach that gives analytics a bad name.

Questions people ask

Is this a psychological test?

No. It is built from decisions the player already made in games, so nobody has to sit through anything and nothing depends on self reporting.

Can we compare a guard with a centre?

Yes for Prediction, which is normalised by role using percentages rather than volumes, so physicality does not decide the comparison.

How many players are covered?

More than 39,500, a narrower set than the full database because the profile needs richer possession data than a boxscore provides.

What does it change for a roster?

It puts a name on the part of fit that used to be intuition, so chemistry becomes something a general manager can plan for rather than discover in November.