# futbolfutbol: science and methods

Version 0.1, 30 October 2026. Written for a club's analysts, sports scientists and general manager.

This document explains how futbolfutbol reads a football match and how we check our work. It describes the methods without the private formulas, and it says plainly what is tested, what is still being validated, and what the system cannot do.

---

## 1. Summary

futbolfutbol combines four kinds of information about a club's games, puts them on one timeline, and uses AI to find patterns, predict what is likely to happen next, and explain it to the staff in their own football language.

- **Inputs:** match data, the club's own data, GPS, video.
- **Common clock:** every source aligned to the same match time.
- **Outputs:** before the game (opponent preview), during the game (live notes), after the game (report and video).

## 2. What goes in

| Input | What it is | Form we accept |
|---|---|---|
| Match data | Events of every game: passes, shots, tackles, with time and position | Football data services (APIs) |
| Club data | The staff's and the organization's own numbers and notes | An internal app, a spreadsheet, paper, a voice note |
| GPS | Distance, speed and load from training and games | XML or CSV (table) |
| Video | The match broadcast or the club's own footage | Video file after the game |

Nothing has to change in how the staff works. If it is written down, recorded or exported, we can use it.

## 3. Data quality

- Before any analysis, the score, shots and key events are re-counted from the raw data and compared across sources. A report is not built until they agree.
- Where two sources disagree (for example, one counts 21 shots and another 23), the report shows both and says which one it uses.
- Every source is placed on one clock: video time, match time and GPS time are aligned, so one moment can be seen in all of them at once.

## 4. The football model

The AI reads every game through the same six levels. The terms at each level are agreed with the club's coaching staff and written as rules, so the AI applies them the same way every time.

1. **Game:** 90 minutes, both teams.
2. **Moment:** attack, defense, and the two transitions.
3. **Zone and lane:** three zones counted from the team with the ball (zone 1 build-up, zone 2 creation, zone 3 final third), crossed with left, center and right. Without the ball, defending in the opponent's zone 1 is a high press, zone 2 a mid block, zone 3 a low block.
4. **Principle:** what the staff wants in that moment and zone (for example, use width and depth).
5. **Play:** a named play the staff recognizes (for example, a diagonal pass in behind to the far-side runner).
6. **Action and rule:** each pass, run or tackle, measured by a written rule (for example: diagonal, forward, ends behind the last defender, a teammate arriving).

## 5. Measures

Standard, published statistical measures, applied the same way to every game. Our own weighting and combinations stay private.

- **Expected goals (xG):** the probability that a shot becomes a goal, from where and how it was taken. It shows whether a result was earned.
- **Pressure over time:** which team is attacking, minute by minute. It shows when a game turns.
- **Pass networks and average positions:** who connects with whom, and where each player really played, half by half.
- **Ratings with a confidence range:** every player rating carries how sure we are, based on how much data stands behind it.
- **Physical load:** distance, sprints and top speed from GPS, set against minutes played and the result.
- **Patterns across games:** the same situation found in many games of a season, so that it can be anticipated.

## 6. The AI

- By default we use **open-weight AI models that we fine-tune for football** and run under our own control.
- Only with the club's written agreement, we also work with **leading commercial AI models through their APIs**, under terms that do not let the provider train on the club's data.
- **The club's data belongs to the club.** It is never sold or shared with another club. **The technology and the science are futbolfutbol's.**
- The models are fine-tuned and enhanced with the club's data and the staff's feedback, so they get better at that club's football the longer we work together.
- The AI does the counting, the matching and the first draft. **A football person reads every page and every drawing** before the club sees it.

## 7. Validation

We test in three ways and publish the results as they come in.

| Test | Status | What it means |
|---|---|---|
| Re-counting | In place | Every report's key numbers are checked against the raw data and a second source. |
| Tests on past games | In place | Models are run on past games and compared with what really happened. |
| Forward validation | Running | Predictions are recorded before games and scored after them, on games the models have not seen. This includes our internal player ranking. First results come with the first partner season. |

## 8. Limits

- Without video or GPS, players off the ball are not seen; positions come from the actions they make.
- Different data services measure some things differently; we show the difference instead of hiding it.
- Patterns need many games. One game is a sample; a season is evidence.
- The system supports the staff's decisions. It does not make them.

## 9. Working with us

We are choosing a small number of partner clubs across the United States, and we talk to clubs abroad. We work privately and in the background, hand in hand with the analysts, the coaches and the general manager, for a full season.

A partner club shares its staff's numbers (in any form), GPS exports (XML or CSV), and match video. What we need most is a steady, consistent feed of that data for a full season: it is what lets the models learn the club. We bring the match data, the AI and the people.

Contact: hola@futbolfutbol.icu · https://futbolfutbol.icu/