24 352
Pronostics checked after the result.
Period: last 120 days · Glicko-2 and math models (no live signals)AI scoreboard
Intelligent AI scoreboard for sports analytics
The AI scoreboard shows automated match analytics, résultat probabilities, live signals and post-result modèle checks.
First we look at strong résultats and the marché's accord with the modèle; exact score remains a reference marché.
Public metric
Pronostic PropickAI accuracy
Available to everyone. Informational. The modèle is calibrated to the marché; this is pronostic accuracy, not betting profitability.
59.24%
Overall hit-rate on the main résultat.
14 426 hits out of 24 352
Model confiance
Average pronostic
Actual (hit-rate)
Pronostics
50–60%
53.0%
58.5%
12 912
60–70%
64.6%
65.2%
4 301
70–80%
74.4%
72.2%
2 277
80%+
84.3%
88.3%
1 260
Calibration: the higher the modèle confiance on the left, the higher the real hit-rate on the right. This means the modèle's probabilities match reality.
Performance by sport
Hit-rate is the share of résultats correctly predicted by the modèle. ROI is the historical return on the modèle's pari de valeurs (flat 1 unit, over the period). Informational, backtest, not a profit guarantee.
Sport
Pronostics
Hit-rate
Value pronostics ROI
Tennis
5 151
65.2%
n/a
Football
4 271
47.0%
n/a
Basketball
1 140
66.3%
n/a
Hockey
364
56.3%
n/a
Volley-ball
277
66.4%
n/a
A hit-rate above the marché does not guarantee positive ROI: at low cotes even high accuracy can go negative (e.g. tennis). Profit holds where the modèle agrees with the marché (basketball).
Method
Comment fonctionne notre IA
PropickAI compares sports statistics, cotes, Glicko-2, live signals and the results of completed matchs. The AI scoreboard shows which models give a stable signal, where the résultat probabilité is confirmed by the marché and how pronostics pass the post-game check.