Scientific Authority
Track Record
How accurate are our signals? Every prediction, verified.
1000 verified predictions across 6 gameweeks. Brier Skill Score: +26.9%
Haul threshold: ≥8 pts (μ=3.2, σ=3.3)
Model Score
0.150
Lower = better
Baseline Score
0.205
Naive xP model
Skill Score
+26.9%
Outperforms baseline
Sample Size
1000
6 GWs
Prediction Accuracy
Points near the diagonal line = well-calibrated predictions
Score Breakdown
How the model score breaks down into calibration, discrimination, and base difficulty
Calibration
0.0390
Discrimination
0.0003
Base Difficulty
0.1109
Where We Were Wrong
Players we expected to haul but didn't. Showing our misses proves our honesty.
P(haul)
78%
Outcome
No haul
P(haul)
75%
Outcome
No haul
P(haul)
84%
Outcome
No haul
P(haul)
84%
Outcome
No haul
P(haul)
83%
Outcome
No haul
Verdict Distribution
937
HOLD
avg Δ: -0.01
5
TRAP
avg Δ: 2.46
58
VOLATILITY_WATCH
avg Δ: 0.36