All Tournaments
Track record Division 1
Every pick is logged before the match starts — before the draft, before anyone has a read on the day. Nothing gets touched after. If I was wrong, it stays wrong, and you can see exactly where and when. That's the only way this means anything. If you can edit your picks after the result, you're not predicting — you're just rewriting history.
Overall accuracy
66%
345 / 519 resolved
Tournaments
12
Total picks
522
Correct
345
Wrong
174
TournamentPicksCorrectWrongPendingAccuracy
Overall456300156—
66%
Track record Division 2
TournamentPicksCorrectWrongPendingAccuracy
Overall6645183
71%
Accuracy counts only resolved picks (correct + wrong). Pending picks excluded from the percentage.
Deep dive
Prediction insights
Patterns across every resolved pick — which teams I read well, when I beat the model, and whether my confidence matches reality.
Best-read teams
Highest accuracy in their series
Hardest to call
Lowest accuracy in their series
Biggest upset calls
Correct picks against the model favourite
51/81
underdog calls landed
Confidence vs. reality
When I claimed a win chance, how often was I right?
- 10–20%n 3low sample33%claimed 16%Δ 18%
- 20–30%n 4low sample100%claimed 24%Δ 76%
- 30–40%n 1283%claimed 36%Δ 48%
- 40–50%n 6258%claimed 46%Δ 12%
- 50–60%n 11562%claimed 55%Δ 7%
- 60–70%n 8972%claimed 65%Δ 7%
- 70–80%n 6674%claimed 74%Δ 0%
- 80–90%n 2886%claimed 84%Δ 2%
- 90–100%n 4low sample75%claimed 94%Δ 19%