Honest accuracy tracking
Accuracy report
Every prediction is stored with a timestamp before kick-off and graded when the result is final. Nothing is deleted or re-scored. Predictions created after kick-off are excluded (56 rows seeded on 2026-06-27, the first day of operation).
AI model leaderboard
Same fixtures, six models
| # | Model | Overall | Match winner | Last 100 | Graded |
|---|---|---|---|---|---|
| 1 | Grok 4.7 xAI | 64% | 63% (178) | 69% | 267/415 |
| 2 | Claude Fable 5.1 Anthropic | 63% | 60% (316) | 66% | 498/793 |
| 3 | Kimi K3 Moonshot | 60% | 57% (2243) | 67% | 3737/6277 |
| 4 | Scorebase Model Elo + Dixon-Coles + market blend | 59% | 56% (3080) | 62% | 5171/8718 |
| 5 | GPT-5.6 Sol OpenAI | 59% | 56% (2828) | 71% | 4606/7780 |
| 6 | Gemini 2.5 Flash Google | 59% | 56% (2420) | 63% | 3172/5406 |
| 7 | Grok 4.1 Fast (retired) xAI | 57% | 56% (2159) | 44% | 3183/5552 |
| 8 | Qwen 2.5 32B (retired) Alibaba (local) | 57% | 56% (1861) | 61% | 2616/4569 |
| 9 | Claude Haiku 4.5 (retired) Anthropic | 57% | 56% (2227) | 58% | 3102/5446 |
Overall = match winner + handicap + over/under combined. Models joined at different dates, so "Last 100" compares them on equal sample size.
Scorebase model by league
Match-winner hit rate
| League | All-time | Last 30d | Last 7d | Strong picks | Graded |
|---|---|---|---|---|---|
| Premier League | 46% | 48% | 60% | 100% (8) | 215/465 |
| LaLiga | 52% | 65% | 80% | 100% (18) | 221/429 |
| Bundesliga | 48% | 47% | 22% | 59% (17) | 159/330 |
| Serie A | 47% | 66% | 59% | 88% (16) | 199/424 |
| Ligue 1 | 48% | 61% | 67% | 100% (5) | 162/338 |
| Champions League | 55% | 67% | — | 83% (6) | 85/154 |
| MLS | 46% | 47% | 52% | 50% (10) | 261/567 |
| K League 1 | 42% | 46% | 33% | 0% (1) | 49/118 |
| NBA | 67% | 64% | 65% | 78% (516) | 945/1418 |
| NHL | 54% | 60% | 68% | 58% (680) | 881/1619 |
| MLB | 55% | 62% | 50% | 56% (1783) | 2004/3650 |
| KBO League | 54% | 64% | 25% | 53% (239) | 497/928 |
| NPB | 56% | 77% | 100% | 59% (433) | 473/839 |
Football draws count as a miss unless the model picked the draw. Strong picks are fixtures where the top probability clears a league-specific threshold.