Open any player report on Vision AI and you'll see a score out of ten, alongside tactical blocks like Breaking Pressure, Progression, Chance Creation, or Ball Winning.

Those numbers aren't designed to tell you how good a player is.

They're designed to answer a much harder question:

What football problems does this player consistently solve?

That difference shapes every part of the evaluation.

Football isn't one skill.

Traditional player ratings usually follow the same formula.

Take a collection of statistics.

Assign each one a weight.

Average everything into one score.

The mathematics may be different.

The football philosophy usually isn't.

But football doesn't work that way.

A centre-back isn't trying to create chances.

A striker isn't expected to dominate aerial clearances.

A holding midfielder isn't judged by the same attacking actions as a winger.

Different positions solve different football problems.

Vision AI starts from those problems—not from the statistics.

From metrics to football questions

Instead of asking:

How many progressive passes?

Vision AI asks:

Can this player move his team forward?

Instead of asking:

How many successful dribbles?

It asks:

Can he escape pressure?

Instead of asking:

How many tackles?

It asks:

Does he consistently stop attacks?

Statistics only exist because they help answer football questions.

Never the other way around.

Every block answers one tactical question

Each player report is divided into independent tactical blocks.

Every block represents a single football question.

Not a statistical category.

A football behaviour.

Examples include:

Breaking Pressure

Can the player protect possession and progress play when his team is genuinely under heavy pressure?

Only actions from high-pressure matches are evaluated.

Not season averages.

Progression

Can he consistently move possession into more dangerous areas?

Whether through carries...

passes...

or both.

Chance Creation

Does he repeatedly create shooting opportunities?

Not simply assists.

Not simply key passes.

The complete creative profile.

Ball Winning

Can he recover possession consistently?

Where does he win it?

How often?

Against what level of opposition?

Box Threat

Does he consistently become a genuine scoring threat?

Not simply by shooting often.

But by arriving in dangerous areas and producing quality chances.

Every block answers one football question.

Nothing more.

Nothing less.

Behaviours—not statistics

One statistic never defines a player.

Progressive Carries don't automatically make someone progressive.

Successful dribbles don't automatically mean press resistance.

Pass accuracy doesn't automatically mean good possession play.

Vision AI combines multiple event-level actions into behavioural profiles.

The objective isn't to describe what happened.

The objective is to describe what the player repeatedly does.

That's why reports describe players through football behaviours such as:

Deep Progressor

Half-space Creator

Wide Delivery Specialist

Box Arriver

Ball-winning Midfielder

Line-breaking Defender

Rather than simply listing statistics.

Every comparison is role-aware

Football positions solve different problems.

A winger isn't compared with a centre-back.

A striker isn't compared with a defensive midfielder.

Every percentile is calculated only against players performing the same tactical role.

A centre-back's progression is compared with other centre-backs.

A winger's crossing is compared with other wingers.

A striker's finishing is compared with other strikers.

Without positional context, statistics lose most of their meaning.

Context changes football

Football actions never happen in isolation.

The same player can produce different numbers simply because he plays in a different environment.

A defender playing for a team with 65% possession receives different situations than one defending deep for ninety minutes.

A winger in a transition team attacks different spaces than one constantly facing a low block.

A striker in a dominant side receives more touches inside the box than one isolated in a defensive system.

Ignoring team context creates misleading conclusions.

That's why Vision AI separates two different evaluations.

Tactical Evaluation

The report score answers one question:

How well did this player solve the football problems expected from his role?

Every tactical block contributes equally.

No hidden weighting.

No tactical preference.

No adjustment for team style.

The report evaluates the player first.

Context-Aware Evaluation

The ranking engine answers a different question:

How impressive is this profile given the environment he plays in?

Here, context matters.

The model considers factors such as:

Possession volume

Territorial dominance

Build-up quality

Pressing intensity

Progression style

Directness

Defensive workload

Attacking support

Carry-heavy versus pass-heavy progression

Wing versus central focus

A creative midfielder in a low-possession side is solving a different football problem than the same player in a dominant possession team.

Likewise, a defender progressing the ball in a deep defensive side is facing different challenges from one playing behind a team that controls every match.

The adjustment is intentionally small.

Context should explain performance.

Not replace it.

Evidence before conclusions

Vision AI never labels a player as creative...

progressive...

press-resistant...

or dominant...

without showing the evidence.

Every tactical block links directly to the actions producing the score.

Every conclusion can be traced back to passes, carries, duels, shots and defensive actions.

The report exposes the reasoning behind every conclusion.

Nothing happens inside a hidden formula.

Confidence matters

Not every number deserves the same level of trust.

A player with only a few appearances shouldn't receive the same confidence as someone with a full season.

Every tactical block carries its own confidence level.

If the player rarely experienced the situations required for evaluation...

the block simply isn't scored.

No guessing.

No artificial estimates.

Small samples produce lower confidence.

Exactly as they should.

What event data cannot see

Vision AI is built on event data.

Event data captures every pass.

Every carry.

Every duel.

Every shot.

Every defensive action.

But it doesn't capture everything.

Without tracking data, it cannot directly observe:

Off-ball movement.

Runs that never receive the ball.

Body orientation.

Scanning behaviour.

First-touch quality.

Pressing angles.

Defensive spacing.

Exact defender proximity.

Rather than inventing signals from missing information...

Vision AI deliberately leaves those questions unanswered.

An honest unknown is always better than a confident guess.

Football is still watched—not just measured

Vision AI isn't designed to replace scouts.

Or coaches.

Or analysts.

It exists to organise evidence.

To separate repeatable behaviours from isolated moments.

To explain why a player succeeds.

To show where context influences performance.

And to make football conversations more objective.

The final decision will always belong to people.

Vision AI simply makes the evidence easier to see.