Prediction Accuracy: July 2026
Every Best Pick and Value Analysis of the month, graded against the final result. Hits and misses count alike; void matches are left out.
Final figures for July 2026
Hit rate in July 2026
Best Picks
The single highest-conviction selection per fixture, across all markets.
+1.7 pts vs June 2026
Value Analysis
Selections where the model's probability diverges most from consensus market pricing.
+1.1 pts vs June 2026
Compared to guessing
An accuracy figure means nothing on its own. The only useful question is what a naive rule would have scored on exactly the same matches.
Measured across every graded fixture, not a selected subset. Home advantage is why the naive rule scores as well as it does; the gap above it is what the model contributes. n = 3,061.
Day by day in July 2026
Each line is a seven-day rolling average of settled selections, with a node marking the days behind it. Days with nothing settled break the line rather than being bridged across.
| Date | Best Picks | Value Analysis |
|---|---|---|
| 01.07.2026 | 67.0% | 63.5% |
| 02.07.2026 | 68.8% | 63.2% |
| 03.07.2026 | 70.4% | 66.0% |
| 04.07.2026 | 69.6% | 62.7% |
| 05.07.2026 | 69.1% | 63.4% |
| 06.07.2026 | 69.2% | 62.7% |
| 07.07.2026 | 70.0% | 63.0% |
| 08.07.2026 | 70.7% | 62.3% |
| 09.07.2026 | 70.6% | 62.5% |
| 10.07.2026 | 68.8% | 62.9% |
| 11.07.2026 | 67.6% | 64.5% |
| 12.07.2026 | 66.5% | 64.3% |
| 13.07.2026 | 66.1% | 64.9% |
| 14.07.2026 | 66.0% | 64.5% |
| 15.07.2026 | 65.0% | 64.6% |
| 16.07.2026 | 65.7% | 63.6% |
| 17.07.2026 | 65.7% | 63.1% |
| 18.07.2026 | 69.5% | 65.2% |
| 19.07.2026 | 69.3% | 63.2% |
| 20.07.2026 | 70.5% | 62.3% |
| 21.07.2026 | 69.6% | 62.3% |
| 22.07.2026 | 68.3% | 61.2% |
| 23.07.2026 | 66.4% | 62.7% |
| 24.07.2026 | 66.6% | 63.3% |
| 25.07.2026 | 62.5% | 61.1% |
| 26.07.2026 | 64.5% | 60.1% |
| 27.07.2026 | 64.8% | 61.2% |
| 28.07.2026 | 65.6% | 61.9% |
| 29.07.2026 | 67.3% | 62.9% |
| 30.07.2026 | 68.1% | 61.0% |
| 31.07.2026 | 68.6% | 60.6% |
Accuracy by competition in July 2026
The competitions with the most graded Best Picks this month, each with at least 20.
| Competition | Best Picks | Value Analysis |
|---|---|---|
 World: UEFA Europa Conference League | 63.6%35 of 55 correct | 62.2%69 of 111 correct |
 USA: USL Championship | 67.5%27 of 40 correct | 57.1%8 of 14 correct |
 Brazil: Serie B | 78.8%26 of 33 correct | 60.9%14 of 23 correct |
| USA: USL League One | 67.7%21 of 31 correct | 75.0%6 of 8 correct |
 Argentina: Primera C | 69.0%20 of 29 correct | 62.5%20 of 32 correct |
| USA: USL League Two | 82.1%23 of 28 correct | 77.8%21 of 27 correct |
| Argentina: Primera Nacional | 75.0%21 of 28 correct | 54.8%17 of 31 correct |
 Ecuador: Liga Pro | 74.1%20 of 27 correct | 50.0%12 of 24 correct |
 Australia: New South Wales NPL | 70.4%19 of 27 correct | 57.9%11 of 19 correct |
| Australia: New South Wales NPL 2 | 70.8%17 of 24 correct | 100.0%5 of 5 correct |
Are the probabilities honest?
This is the test that matters. We group every match-result forecast by the probability we published, then plot what actually happened. A calibrated model sits on the diagonal.
What this chart coversMatch result across every graded fixture, not just selected picks - a selected subset could not test calibration honestly. Buckets holding too few fixtures to plot are merged into the end points, and point size scales with sample. Vertical bars are 95% intervals: a thin bucket is an uncertain reading, not a precise miss.
Brier score: 0.2003Mean squared error between published probability and outcome. Lower is better; 0.25 is what you would score by predicting 50% on everything.
Calibration by probability bin
| Probability band | Probability we published | How often it actually happened | 95% interval | Fixtures |
|---|---|---|---|---|
| 0.05-0.20 | 14.7% | 16.5% | 14.6% – 18.5% | 1361 |
| 0.20-0.25 | 22.6% | 21.8% | 19.8% – 23.9% | 1530 |
| 0.25-0.30 | 27.4% | 27.6% | 25.6% – 29.6% | 1921 |
| 0.30-0.35 | 32.4% | 29.6% | 26.9% – 32.5% | 1032 |
| 0.35-0.40 | 37.3% | 34.7% | 31.4% – 38.1% | 778 |
| 0.40-0.45 | 42.4% | 46.4% | 42.6% – 50.2% | 658 |
| 0.45-0.50 | 47.4% | 47.2% | 43.1% – 51.3% | 564 |
| 0.50-0.55 | 52.4% | 54.8% | 50.1% – 59.4% | 438 |
| 0.55-0.60 | 57.3% | 60.0% | 54.5% – 65.3% | 315 |
| 0.60-0.65 | 62.4% | 62.4% | 55.8% – 68.7% | 213 |
| 0.65-0.70 | 67.4% | 67.6% | 59.7% – 74.6% | 148 |
| 0.70-0.75 | 72.6% | 69.9% | 58.6% – 79.2% | 73 |
| 0.75-0.80 | 77.2% | 74.0% | 63.3% – 82.5% | 77 |
| 0.80-0.95 | 85.7% | 76.0% | 65.2% – 84.2% | 75 |
Interval is the 95% range the observed figure could take at that sample size. Narrower means more matches in the bin.
Which markets the picks came from
A hit rate across a mixed basket is not the same claim as a hit rate on match result alone. These are the six most-used markets in each basket, with the remainder grouped, so you can judge both headline figures for yourself.
Best Picks
- Home Win or Draw36.9% of picksn = 442
- Under 3.5 Goals25.6% of picksn = 307
- Draw or Away Win15.3% of picksn = 184
- Home Win13.3% of picksn = 159
- Away Win4.3% of picksn = 52
- Over 1.5 Goals3.1% of picksn = 37
- Over 2.5 Goals1.5% of picksn = 18
Value Analysis
- Under 2.5 Goals22.0% of picksn = 286
- Under 3.5 Goals17.8% of picksn = 231
- Over 1.5 Goals10.3% of picksn = 134
- Home Win or Draw10.2% of picksn = 132
- Over 2.5 Goals7.2% of picksn = 94
- Home Win6.6% of picksn = 86
- Other markets (27)25.9% of picksn = 336
3 void matches were excluded from these figures.
Monthly reports
How the predictions did in July 2026
813 of 1199 graded Best Picks were correct in July 2026, a hit rate of 67.8%. Value Analysis was right in 810 of 1299 matches (62.4%). That is 1.7 points more than in June 2026.
On match result, the model was right in 51.9% of 3061 matches, against 44.7% for always picking the home team. Among competitions with at least 20 graded Best Picks, USL League Two had the highest hit rate: 82.1% of 28.
Every prediction is graded against the result after 90 minutes, hits and misses alike, and void matches are left out. The all-time record and the method are on the Accuracy page.