Altora Analysis

🎯 DARTS — ACCURACY & RECORD

Calibrated model analysis, not tips. Validated out-of-sample, tracked in the open.

← all sports

Today · 2026-09-04

Today's model reads

Model win probability for today's matches — not tips. These settle into the record below.

Luke Woodhouse v Connor Scutt
65% · mkt 63%
Krzysztof Ratajski v Adam Gawlas
66% · mkt 67%
Dirk van Duijvenbode v Mario Vandenbogaerde
68% · mkt 68%
Martin Schindler v Robert Owen
57% · mkt 58%
Beau Greaves v Daryl Gurney
60% · mkt 59%
Rob Cross v Keane Barry
68% · mkt 68%
Damon Heta v Karel Sedláček
58% · mkt 59%
Niels Zonneveld v Jeffrey de Graaf
50% · mkt 50%

Model updated 22 July 2026: published probabilities now blend the Elo rating with the market opening price through a gradient-boosted calibration layer, validated out-of-sample before switching. The record below spans both model versions.

Model validation · 3,373 out-of-sample matches · to 2026-07-26

Refreshed 26 July 2026 under a documented, reproducible protocol: the serving form + rating blend, priced with each competition's real match format, tested on 2024-onward matches. It replaces an earlier snapshot whose producing script was lost — we rebuilt the harness rather than carry numbers we could no longer verify.

Is the model accurate?

These are matches the model never trained on. Two things prove accuracy: it picks the winner about two in three, and — more importantly — its probabilities are calibrated: when it says 71%, it happens 71% of the time.

62%favourite pick accuracy
0.227model brier
0.223market brier

Reliability — model probability predicted vs what actually happened

30%
28%
242
44%
44%
915
57%
57%
1,404
71%
70%
720
84%
96%
84

The two bars line up on every band — that's calibration. Not sharper than the market (that's near-impossible), but demonstrably accurate, and shown in the open.

The live record · every pick since going public

The live record

152settled
69%model pick hit
0.209model brier
0.212market brier

Nothing deleted — wins and losses, every one. The live sample is young; the validation above is the model's accuracy over thousands of matches.

DateModel pickProbResult
6 voided match(es) shown greyed below — kept on the record, excluded from the accuracy figures (no result to score).
2026-09-04Cameron Menzies v Johan Engstrom83%W
2026-09-04Andrew Gilding v David Rosi90%W
2026-09-04Kim Huybrechts v Ryan Joyce62%L
2026-09-04William O'Connor v Jiri Brejcha88%W
2026-09-04Niko Springer v Martin Homola86%W
2026-09-04Kevin Doets v Ricky Evans66%W
2026-09-04Joe Cullen v Stephen Burton67%L
2026-09-04Cristo Reyes v Callan Rydz59%L
2026-08-30Rob Cross v Dirk van Duijvenbode60%L
2026-08-30Ryan Searle v Danny Noppert54%L
2026-08-30Jonny Clayton v Ross Smith56%L
2026-08-30Keane Barry v Dave Chisnall52%L
2026-08-30Gian van Veen v Jermaine Wattimena60%W
2026-08-30Gary Anderson v Stephen Bunting66%W
2026-08-30Nathan Aspinall v Michael van Gerwen55%W
2026-08-30Chris Dobey v Josh Rock56%W
2026-08-30Dirk van Duijvenbode v Danny Noppert54%L
2026-08-30Ross Smith v Dave Chisnall68%W
2026-08-30Gary Anderson v Gian van Veen56%W
2026-08-30Nathan Aspinall v Chris Dobey52%W
2026-08-29Ross Smith v William O'Connor63%W
2026-08-29Wessel Nijman v Dave Chisnall74%L
2026-08-29Jermaine Wattimena v Andrew Gilding57%W
2026-08-29Danny Noppert v Ritchie Edhouse59%W
2026-08-29Ryan Searle v Niko Springer55%W
2026-08-29Rob Cross v Luke Woodhouse60%W
2026-08-29Gary Anderson v Krzysztof Ratajski57%W
2026-08-29Chris Dobey v Tom Sykes63%W
2026-08-29Stephen Bunting v Jonas Masalin82%W
2026-08-29Jonny Clayton v Laszlo Kadar91%W

⬇ Download the full record (CSV — opens in Excel / Sheets) · showing the most recent 30 of 152