The same numbers your statistician trusts
The modules were carried over from their tool one to one and are guarded by tests that compare the result cell by cell — quirks of the old program included. Changing the tool does not change the numbers.
From the court to the conclusion.
You win games. We show you why.
You record the game possession by possession, the stats compute themselves, the report says what happened, and the system points at specific training drills — on your data, on your server.
- 13
- statistical modules
- 9 × 5×5 and 4 × 3×3, matching the statistician's sheet
- 610,045
- recorded actions
- across the 4129 games already in the system
- 12
- training rules
- each with a named threshold and a measure of certainty
- 142
- training drills
- straight from the coaching literature, with criteria and a source
Four steps, one chain of data
Nothing has to be retyped between stages. What the person at the scorer's table writes down reaches the training recommendation.
- 1
You record the game
Possession by possession: action code, shot zone, time, the five on court, the player finishing it. From a phone or a laptop at the table, with nothing to install.
- 2
The stats compute themselves
Thirteen 5×5 and 3×3 modules — the same tables your statistician has been building by hand, matching their sheet cell by cell.
- 3
The report says what happened
Wins against losses, the rhythm of the quarters, runs, shot zones, the on-court balance of each five, a player's trend across the season. With a warning when the sample is too small for a conclusion.
- 4
The system says what to practise
Twelve rules ask your data about specific weaknesses and point at drills from the coaching literature. Every recommendation shows the numbers it was built on.
This is what it looks like inside
Screenshots from a running instance, on real games from professional competitions — not mockups. Pick an item from the list or wait for it to advance on its own.
What you get
Six things that together close the path from recording a game to deciding on a training session.
A report that says what happened
A comparison of any two sets of games, the rhythm of the quarters, runs, a map of shot zones, the on-court balance of each five. Ready to print and to send.
From “what to improve” to “what to practise”
You record a focus for a player, and after a few games the system checks by itself whether it improved. Plus a catalogue of drills matched to what the data shows.
A system that can say “I don't know”
Every verdict comes with a confidence interval. When the sample is too small you get “check this”, not an invented diagnosis. The audit screen shows every rule — including the ones that stayed silent.
You ask in plain language
The assistant in the app answers questions about your games and shows which data it used. Every number in the answer comes from the database — tests make sure the model does not invent one.
Your data, your server
The platform runs on your own infrastructure and the club's data stays in the club. Each organization sees only its own games.
Who we are talking to
The same platform answers different questions for a coach, a statistician and a board. Pick yours.
Coach
From the game to the training decision, without a spreadsheet.
See the details →
Club statistician
Your modules cell by cell, sheet import and export, a file validator.
See the details →
Club board
One platform for every team, data on your server, a trace of every change.
See the details →
Player
Your real contribution and your trend across the season, not just points.
See the details →
Parent
A readable summary of the game and a way to ask about it in plain language.
See the details →
Fan
How your club and its rival play, on data rather than impressions.
See the details →
What we do not promise
So there are no surprises at the second meeting.
- There is no price list and no self-service purchase — a rollout is agreed individually.
- The 3×3 modules are ready, but there are no real 3×3 games in the system yet.
- The report narrative is written by a rule engine, not by a language model.
- Recognising actions from video does not exist yet — today a human records the game. It is on the roadmap and we will say so plainly once it works.