05Family 02 of 05

The map

Where does the evidence come from? Coordinates alone, pixels alone, or every witness reconciled into one match state. It is the first decision in any possession system, and every later decision inherits it.

Your question The study that answers it Three witnesses, one state? The single source of truth Coordinates alone? Geospatial-only inference Pixels alone? Video-only inference
The decision map of this family. Each question leads to the study that answers it.
  • 3studies in this family
  • 3sources of evidence
  • 2engines any of them can feed

One instant, three witnesses

The choice in this family starts from what each kind of witness can and cannot see at one moment of one match. Everything else follows from those blind spots.

The same instant three times. Tactical tracking has everyone and no ball. Broadcast tracking has the ball inside a window. The video frame has contact and a shirt number, and no metres. Below: where each study of this family sits between one witness and one reconciled state.

Every route on one page

A possession system is a path through two decisions: where the evidence comes from, and how evidence on single frames becomes one sequence. Any source can feed any engine, so there are more routes than boxes. Laying them all out first is what keeps a team from building the fashionable one.

Three sources, two engines, one sequence. The lit route is the one we fielded: tracking into evidence and a decoder.

The sources, side by side

Criterion Tracking only Video only All reconciled
Spatially normalised on arrival Yes No, needs calibration After calibration
Survives an invisible ball With ball-free reasoning Poorly Yes, where another witness sees it
Survives unstable identities With order-free features Re-identification is hard Shirt numbers from pixels
Needs synchronisation No If two cameras are used Yes, per segment
Explainability High for physical rules Variable Depends on provenance
Complexity Lowest High Highest

How an architect chooses here

  • Start from what the sensor delivers. If tracking exists, coordinates are evidence you can audit today.
  • Count what you would have to own. Detection, calibration and synchronisation are products in themselves.
  • Add a witness for its different failures. Two sources that fail together are one source.

What this family cannot do

A source gives evidence about single instants. It does not give a sequence. Turning 25 opinions a second into one story of the half is the job of the engine.

In this family3

  • Three witnesses watch one match: the video frames, the tactical tracking and the broadcast tracking. Each has gaps the others can fill. This is the architecture that reconciles ...

    • Three witnesses
    • Fusion
    • Provenance
  • Possession from coordinates alone: player and ball positions, speeds and tactical tags. No pixels at inference. The question is which reasoning to put on top.

    • Tracking
    • Tags
    • Laptop CPU
    • Fielded
  • Possession from pixels: detect the ball and players, calibrate the pitch, project to the ground, then reason. Independent of any supplied tracking, and expensive at every step.

    • Computer vision
    • Detection
    • Calibration