Model Performance
Radical transparency: see exactly how accurate our predictions have been.
Loading performance data...
Why This Matters▾
Real games, real results
Every prediction is graded against the actual final score. Nothing is simulated or back-tested.
Calibration
When we say 60%, it should happen ~60% of the time. Not cherry-picked hindsight.
Full Transparency
We show our weak spots too. No model is perfect—ours is honest about its limits.
Performance Benchmarks▾
How trust is graded
- • Strong — model is consistently right, well above chance
- • Good — beating chance, small but real edge
- • Borderline — close to chance, treat with caution
- • Unreliable — below chance, avoid following
Context Matters
- • Football is high‑variance — upsets happen
- • 1X2 random baseline is 1 in 3 (33%)
- • Goals and BTTS random baseline is 50%
- • A small sample of results will be noisy — give it time
Try Our Predictions
See probabilities backed by transparent, validated performance