Bet Finder
Find the markets you want to research. Compare match estimates with stored odds, check the evidence and follow the results.
Stored research sample — freshness details
Sample read started 2026-11-07T23:10:42+00:00. Complete samples are shared for at most 15 seconds; incomplete samples for at most 5 seconds. Original model and odds timestamps are unchanged; age and kickoff checks are repeated for each view.
Reads use stored data, not live bookmaker quotes. Fixture, score, model and current-price records from one read-only database snapshot. Price-change history and competition totals are separate reads; records and prices can change after a sample was taken.
Choose a marketFilter goals, match results, BTTS or double chance.
Compare the evidenceCheck the estimate, price and data timestamp.
Follow it throughOpen the match for context or review captured selections in the record.
What are you looking for?
Custom filters ↓Choose a shortcut, then refine the competition and minimum estimate below.
Upcoming fixtures first, then live and finished research. Model estimates and bookmaker-derived market chances are labelled separately. Delayed, postponed and cancelled records appear last.
Probability or value? Probability describes how likely an outcome is. Value also considers its price. A high percentage alone does not make a good selection; estimates and odds can change.
About these probabilities and odds
Kickoff Edge combines recorded expected goals (xG) with opponent-adjusted team strengths. The xG model blends the recent xG signal and v3 equally, using up to ten previous matches per team, weighted toward recent performances. Both teams need at least five xG matches within six months. Where xG is unavailable, the labelled team-strength model remains in use, with recent-form fallback for limited competition history. Experimental recent-form fallbacks remain research only and are excluded from probability and value recommendation filters. Missing xG is never treated as zero. Competitive cup matches between verified domestic divisions also adjust team strength for the division gap, using earlier league appearances. The initial map covers 37 leagues in eight national pyramids; unknown or cross-country tiers are not guessed. All markets share one score distribution. These are pre-match estimates, not live forecasts or guaranteed outcomes. Historical testing showed lower overall error on the tested sample, but calibration and future performance remain under review. Odds show the highest stored price for each selection; bookmaker availability and prices can change. Additional xG and multi-source challengers are evaluated separately in the model observatory before public promotion. Bookmaker price movement is recorded from collection onwards; it does not reveal betting volume.
No safely loaded fixtures match these filters
Try another day or remove the minimum probability filter. Missing model data is not treated as zero.
Reset filters →