Bundesliga prediction accuracy

Every forecast is locked before kickoff (normally 24 hours before), hashed into our public ledger and graded on the 90-minute result. Misses stay on the record. Published forecasts and the backtest are reported separately and never mixed.

Data batch updated Early sample: 3 graded
RPS0.152Not enough rounds for an interval
Log loss1.133Lower is better
Picks graded331 Oct – 1 Nov 2026
Correct (1X2)33.3%Most likely result came true
Brier score0.705Sum over H/D/A, 0–2
Coverage100.0%0 awaiting grading · 0 excluded

Model comparison

ModelnCorrectRPSBrierLog loss
Final (published)V1: equals the statistical model333.3%0.15200.70521.1329
Statistical modelElo + Dixon-Coles333.3%0.15200.70521.1329
ClaudeAI analyst · shadow1100.0%0.12130.30500.5978
CodexAI analyst · shadow1100.0%0.11050.27860.5621
Base rateBenchmark: league H/D/A frequency333.3%0.16290.72171.1904
Elo onlyBenchmark333.3%0.14640.67781.0820

The n column shows how many matches each row covers, so only compare scores between rows with the same n. Lowest RPS among the rows with the most matches is in bold. Claude and Codex have only run on some matches so far and stay in shadow: they are graded but not used in the published probability (V1). The AI layer section compares them with the statistical model on the same matches.

Calibration

Graded published forecasts in five bins by the top probability, with 95% intervals.

30%30%40%40%50%50%60%60%70%70%80%80%90%90%45–55%: predicted 46%, observed 0% (n = 1)55–65%: predicted 60%, observed 50% (n = 2)Predicted probability of the pick

Our picksPerfect calibration

Top probabilitynPredictedObserved95% CI
< 45%0–––
45–55%146%0%0–79%
55–65%260%50%9–91%
65–75%0–––
≥ 75%0–––

By confidence tier

TiernCorrectBacktest
High0–70–83%
Medium250.0%about 61%
Low10.0%39–49%

High: top probability 65% or more · Medium: 50–65% · Low: under 50%. Totals include Low picks.

By market

"Simple rule" is the benchmark in the note under each market.

MarketCorrectSimple rule
Result (1X2)Always pick the home team33.3%33.3%
Correct score: most likely overallExact scores land about one time in eight33.3%–
Correct score: most likely in the pick0.0%–
Over/Under 2.5 goalsAlways pick the more common side66.7%66.7%
Both teams to scoreClose to the base rate: little skill shown so far100.0%100.0%

AI layer – shadow, not yet used in published probabilities

On the 1 graded match where every layer has a forecast.

LayerCorrectRPSLog lossRPS vs statistical
Statistical model (published)100.0%0.10920.5570–
Claude100.0%0.12130.5978+0.0121
Codex100.0%0.11050.5621+0.0013
AI layer combined (shadow)100.0%0.11050.5615+0.0013

Conditions to use the AI layer in published probabilities (all leagues together)

  • At least 300 graded matches and 6 weeks: 11 graded with the AI layer over 0.1 weeks
  • AI layer beats the statistical model on log loss and RPS, on the same matches: AI layer minus statistical: RPS +0.0003, log loss +0.0009 (negative is better; the next condition checks whether the gap is real)
  • 95% interval of the RPS difference excludes zero: -0.0004 to +0.0010
  • Improvement holds in at least two-thirds of leagues: 1 of 4 leagues with AI forecasts

Not all conditions are met, so the AI layer stays in shadow. Negative RPS difference = better than the statistical model.

Grading rules

Graded on the 90-minute score; extra time and penalties are ignored. Every forecast locked before kickoff counts. We lock 24 hours before kickoff, or straight away when a match enters our tracking later than that; each match page shows the actual lock time. Forecasts locked after kickoff are never graded. Corrections are logged, never silently replaced.

Grading methodology

Exclusion log

No exclusions so far.

Forecast archive

Every graded forecast with its locked probabilities, result and hash. Download the full ledger as CSV or JSON.

Browse graded resultsOpen the ledger

Other leagues: Premier LeagueLa LigaSerie ALigue 1Primeira LigaScottish PremiershipEredivisieSüper LigEFL ChampionshipSaudi Pro LeagueChampions LeagueEuropa LeagueConference League All leagues

For information and entertainment only. Not betting advice. 18+. Predictions are probabilities, not certainties. Responsible gambling