AI Insights
Every flag in the product right now is a threshold rule you set and can trace back to a number. Anomaly detection and injury risk forecasting are on the roadmap, built on the load and wellness history you're collecting now. Neither one is in the product today.
What runs now
Every flag in Fractall today is a threshold rule on ACWR, strain or a wellness score. You set the boundaries and you can trace exactly why an athlete tripped one. It's a blunter instrument than a model, and it has the advantage of being auditable.
ACWR flags
Undertraining, risk and overtraining, against boundaries you set.
Strain flags
High and critical, on the monotony and strain calculations.
Wellness bands
Excellent through very low, on the 5 to 25 composite score.
Explainable
Every flag traces to a number and a threshold.
What's planned
The data to build on is accumulating now, across load, wellness, body pain and injury records. Prediction comes after there's enough of it to be worth trusting.
Anomaly detection
Surfacing the athlete whose pattern broke, rather than the one who crossed a line.
Injury risk forecasting
Built on the load and wellness history plus logged injuries.
Timing
On the roadmap. We'll quote a date when we have one worth quoting.
Straight answer
Injury prediction gets marketed hard in this category. Fractall doesn't have it, and calling threshold flags a prediction engine would be a stretch we're not making.
Not built
No predictive model, no anomaly detection, in any form.
Not implied
The flags are rules, and the page says so.
Want to know when predictions ship?
Tell us what you'd want flagged and we'll come to you when there's something to test.