Risk Advisor weights

TL;DR

The Weights sub-page analyzes which indicator was red in the 14 days before each logged injury in the org, and suggests raising or lowering each indicator’s weight for future versions of the score. The page is informational only today — automatic weight adjustment will be applied in a future iteration. Available on Pro and Enterprise.

For sales

• Problem: Risk Advisor weights are the same set for everyone (Williams 2017 + Drew/Finch 2016). But injury patterns vary by average age, competition intensity, and position. It must be possible to calibrate per club.
• Use cases: quarterly review by the HoP to decide whether the score’s sensitivity is right for their squad.
• Plans: Pro and Enterprise only. Essential sees a banner to view plans with a link to Upgrade.

How staff use it

Access

Pro and Enterprise plans only. Essential sees a banner pointing to plans.

Step-by-step flows

• Top stat strip — 3 KPIs: number of analyzed injuries, number of pre-injury snapshots cross-referenced, analysis window (14 days).
• Insufficient sample banner — if there are fewer than 10 injuries logged, a yellow banner reads “suggestions are indicative only”.
• Main table — one row per indicator (today 4: ACWR distance, ACWR HSR, Hooper, pain), columns:
  • Pre-injury red freq. — % of snapshots in the 14 prior days where that indicator was red.
  • Current weight — what is applied today in the calculation.
  • Suggested weight — internal heuristic: if an indicator was red frequently before injuries it suggests raising its weight; if almost never, lowering it, with caps to avoid extreme values. The exact factors are managed inside the app.
  • Action — badge ”↑ Raise weight” / ”↓ Lower weight” / “No change” / “No data”.

Related configuration

• Current weights are fixed for all orgs today. There is no UI to apply suggested weights automatically — the workflow is: read the page and request the change from support.
• Thresholds (which define when an indicator is red) are configurable per org in Settings → Risks.

FAQ / edge cases

• “No data” on an indicator: no snapshot in the pre-injury window had a value for that indicator. Common with Vmax% (requires ≥5 sessions) or with new orgs without GPS data yet.
• Why only 4 indicators? The other 6 in the score (Vmax%, ACWR sprint, ACWR HMLD, injuries 6m, medical not-fit, physio, anxiety) are boolean or strongly correlated with the first ones — calibration analysis adds nothing new today. Can be expanded after one iteration.

How the player sees it

Player surface: N/A. Staff-only and Pro/Enterprise only.

Data and metrics

Frequency calculation

For each injury with a recorded date in the org (max 500 most recent):

1. Fetch all risk snapshots for that player.
2. Filter snapshots dated in the 14 days prior to the injury.
3. Accumulate how many snapshots had the indicator red and the total with a value.
4. Red frequency is the division of both.

Suggestion heuristic

The suggestion compares how often each indicator was red in the pre-injury window against its current weight: high frequency → suggests raising the weight, low frequency → suggests lowering it, in between → no change. Upper and lower caps prevent extreme weights even with small samples. The exact thresholds and factors are managed inside the app.

Integrations

• Main Risk Advisor — suggested weights refer to the score calculation. Per-org weight overrides will be enabled later.

Limitations

• Does not apply suggested weights automatically — observability only. Applying them requires human decision until A/B validation exists.
• Recommended minimum sample of 10 injuries; below that the banner warns it is indicative.
• Analyzes only 4 indicators; the rest are evaluated qualitatively.