Model-driven analysis across sports and markets, tracked in the open
Every prediction published before the event.
Every result tracked forever.
Nothing deleted.
Free daily predictions for every sport we model, on the site and on Telegram — football, basketball & handball picks go out at the opening price. An independent, free project: no paid tier, nothing held back.
| Sport | Result data | Matches | Competitions | Span |
|---|---|---|---|---|
| Esports (per-title) | PandaScore | 207,132 | — | 2014–2026 |
| Basketball | API | 175,374 | 338 | 2017–2026 |
| Handball | API | 128,868 | 179 | 2016–2026 |
| Tennis | Tennis-data | 103,520 | — | 2005–2026 |
| Football | API | 72,410 | — | — |
| Esports | Betfair | 62,459 | — | 2017–2026 |
| Darts | Betfair | 41,420 | — | 2016–2026 |
| Rugby Union | API | 40,111 | 74 | 2008–2026 |
| Snooker | Betfair | 22,482 | — | 2017–2026 |
| Rugby League | Betfair | 4,535 | — | 2017–2026 |
| Total | 858,311 | 591 | since 2005 |
Matches in our historical database, counted from the live data files on every deploy. API = structured results with scores and stable team ids; Betfair = results reconstructed from settled markets, so no scores; Tennis-data = tennis-data.co.uk archives with results, scores and closing odds; PandaScore = per-title esports match histories with map-level results, powering the dedicated CS2 and Dota 2 rating models. That difference is why some models are stronger than others, and we publish it rather than average it away.
Read the methodology → the models, the data, our exact thresholds, and how to verify the record yourself.
The Handbooks
The complete method behind every record on this site, written for beginners, with full Python code at every step and the honest testing most guides skip.
Educational guides, not betting or financial advice · £29.99 each · instant PDF via Whop
Questions about the models, the record, or Altora? Drop us a line.