Platform core above every product
President Hub The whole club on one board screen Ask about world football or your club Players Worldwide 250,000+ players, free to search
Sporting side · 9 hubs
Marketplace for clubs
Talent market
Also from TSH AI Football DecisionScout Radar
Sporting Director HubScouting HubAnalytics HubCoaches HubPerformance HubAcademy HubMedical HubPlayers MarketplacePlayer App On every pass we see all the options the player had and show whether he chose the best one. Stats see the outcome. We rate the decision.
By invitation · Premier League and Eredivisie 2025/26
Chose the wrong pass
A 4.7× more valuable option was on.
Best option on the pitch
Only 9% of players pick this pass.
Real rated pass · Premier League 2025/26 · players anonymised. Ring = an option (size: how often players pick it, colour: its value) · solid arrow = pass played · dashed = best option.
At the moment of the pass it sees every teammate the ball could have gone to, values each option and compares it with what the player actually did. Coaches spot poor decisions without hours of video; scouts see who makes the right call more often than anyone else in his position. Live today for the Premier League and Eredivisie, by invitation.
Who it is for
Coaching staff
Match reviews and opposition prep without hours of video
Analysts
Key moments and player profiles ready to use, instead of clipping video by hand
Scouts and sporting directors
See how a player makes decisions before you sign him
Agencies
Solid, measured arguments for your player in negotiations
Stats see the outcome. We rate the decision.
We see every option
At the moment of the pass the model knows where every visible player was and sees every teammate the ball could have gone to.
We rate each option
For each one it estimates how likely the pass is to be completed and how much closer it would take the team to a goal.
We compare it with what he did
If there was a clearly better option and the player went elsewhere, that's a mistake, even if the pass was completed. A strong pass few players would see is a plus.
We split decision from execution
Every pass is broken into situation, decision and execution. A player can pick the wrong option and still play it accurately, or the other way round. Standard stats lump the two together.
The pass was completed, so standard stats count it as a plus. The decision was poor: a better option was on.
How good the position was when he got the ball
Did he pick the best available option?
How accurately he carried it out
The pass as a whole: what outcome-based stats see
The key moments of every match, already picked out. Where players made the wrong decision and where they got it right, each with the minute and a pitch freeze-frame: rings on every option, the pass played and the best option side by side.
Options at the moment of the pass
| Value | Completed | Picked by |
|---|---|---|
| Best | 85% | 29% |
| 69% | 4% | |
| 79% | 9% | |
| Played | 93% | 10% |
| 98% | 25% |
← → steps through every !! and ?? of the match.
Who sees it, who executes it, who has work to do. A squad scatter of decision quality against execution over every player in the league, plus each player's typical mistakes and strong decisions compared with his role.
Who to press, and where their decisions go wrong. Your next opponent's weakest decision-makers, the team's error profile against the league and a pitch-zone heatmap of their mistakes, ready for the pre-match meeting.
Under pressure they pass backwards 13.9 times in 100 · league average 12.3Pressing habit · stable across the season: r 0.60
Game intelligence, measured rather than judged by eye. Any player in a covered league, compared on decision-making with everyone in the same position. Filter by league, role and minimum passes, then read the results on the chart.
| Player | Decision | Execution |
|---|---|---|
| 1 | 100 | 92 |
| 2 | 99 | 100 |
| 3 | 97 | 10 |
| 4 | 96 | 63 |
| 5 | 94 | 54 |
Percentiles among Premier League central midfielders · names hidden on this page
A good decision moves the attack towards goal. Standard stats count a safe pass back as a plus and a brave ball into the box that doesn't come off as a minus. We rate what the player chose against what was on. It runs on event data with the positions of every visible player at the moment of the pass.
Six decision-error types
Four strong-decision types
1.4×
Decision-making is a skill
Our decision rating is 1.4× more consistent from match to match than standard passing stats. Tested on matches the model had never seen.
0.92
Predicts the pass before it is played
It predicts whether a pass will be completed before it is played, scoring 0.92 out of 1, ahead of standard models that already know where the ball went.
48%
Knows who gets the ball
Among six or seven teammates it picks the actual receiver 48% of the time, against 33% for 'pass to the nearest teammate'.
D ≠ E
Two separate skills
Decision and execution barely move together, so a player's passing accuracy says little about whether he picks the right pass. We show both.
No. The analyst gets the key moments already picked out and spends the time drawing conclusions, not clipping video.
We license event data with the positions of every visible player at the moment of the pass from specialist providers. Later we'll add our own video tracking.
Live today for the Premier League and Eredivisie, season 2025/26. The method is validated in seven European leagues. Adding a league needs no new development: we plug in the data and the league appears in the product.
AI Football Decision is live by invitation. Request access and we'll set you up with your league and your club.
Long switches to the far flank aren't rated yet; they need full-pitch tracking. Possession-dominant teams are measured against the league norm, so their numbers read slightly lower; a benchmark against the team's own game plan is planned.
From 300 rated passes. Below that we still show his moments, but the rating is flagged as a small sample.
Request access and we'll set you up with your league, your club and your next opponent.
