Experimental AI-assisted analytics — predictions and ratings are estimates, not official results.
predictFRC
Transparency

Data Sources & Attribution

predictFRC combines official competition data with its own experimental analytics. Every important value carries a source label so you always know whether you're looking at an official result or a APEX estimate.

This project is independent and is not affiliated with FIRST, Statbotics, The Blue Alliance, or any official FRC organization.

Source labels

FIRST officialThe Blue AlliancepredictFRC calculatedEstimatedAI-inferredUser-submittedMock data

Source of truth by category

CategoryAuthoritative order
Team identity & metadata
FIRST→TBA→MOCK
Event schedules & dates
FIRST→TBA→MOCK
Official rankings
FIRST→TBA→MOCK
Match results
FIRST→TBA→MOCK
Score breakdowns
FIRST→TBA→MOCK
Alliance selections
FIRST→TBA→MOCK
Awards
FIRST→TBA→MOCK
Match videos
TBA
APEX ratings
APEX
Win probabilities
APEX
Match predictions
APEX
Confidence scores
APEX
Team contribution estimates
EST
Detailed scouting metrics
USER→AI→EST

How source conflicts are handled

  • • Both raw values are preserved — official scores and rankings are never averaged.
  • • The configured authoritative source is selected; the other is recorded.
  • • Official data is never overwritten by AI-inferred or estimated data.
  • • A temporary source delay is not treated as an error until a configurable freshness window passes; only then is it flagged as a conflict.
  • • When a disagreement affects a visible metric, a data-quality warning is shown.

2026 REBUILT game data

The 2026 REBUILT game configuration is currently version 1.0.0 with 2 scoring constants still flagged unverified. Point values are placeholders pending field-by-field reconciliation against the official 2026 Game Manual and Team Updates. The scoring engine is config-driven and unit-tested, so verified values can be dropped in without code changes.

Refresh timing

  • • Team identity & historical data: cached up to 24 hours.
  • • Event metadata: cached up to 6 hours.
  • • Upcoming schedules: refreshed about every 30 minutes.
  • • Active-event results & rankings: refreshed about every minute.
  • • Failed or empty responses never erase good cached data (stale-while-error).

Report an error

Spotted something wrong? Contact [contact email]. Official FIRST / FMS results always take priority over APEX estimates.