Six from nine. Season holds at 71.4%.
Results
6/9 correct. The misses covered the range — a low-confidence call gone wrong (Western Bulldogs, 54%), a low-confidence call gone wrong the other way (Gold Coast, 56%), and a high-confidence call that blew up (Geelong, 79% at Adelaide Oval, lost by 30). Melbourne were rated 79% favourites and won comfortably, which is the model working as intended for once.
Technical
| Match | Predicted | Prob | Result | Margin |
|---|---|---|---|---|
| Western Bulldogs vs Sydney | Western Bulldogs | 54% | ✗ | −66 |
| Richmond vs Melbourne | Melbourne | 79% | ✓ | −54 |
| Hawthorn vs Gold Coast | Gold Coast | 56% | ✗ | +49 |
| Essendon vs Collingwood | Collingwood | 71% | ✓ | −77 |
| Port Adelaide vs Geelong | Geelong | 79% | ✗ | +30 |
| Fremantle vs Carlton | Fremantle | 81% | ✓ | +14 |
| St Kilda vs West Coast | St Kilda | 80% | ✓ | +101 |
| Brisbane Lions vs Adelaide | Brisbane Lions | 80% | ✓ | +52 |
| GWS vs North Melbourne | GWS | 64% | ✓ | +7 |
Season: 45/63 (71%)
| Round | Correct | Total | Accuracy | Model |
|---|---|---|---|---|
| Opening Round | 3 | 5 | 60% | Logistic |
| Round 1 | 7 | 9 | 78% | Logistic |
| Round 2 | 6 | 7 | 86% | Logistic |
| Round 3 | 4 | 7 | 57% | Logistic |
| Round 4 | 5 | 8 | 63% | Logistic |
| Round 5 | 6 | 9 | 67% | Logistic |
| Round 6 | 8 | 9 | 89% | Phase NN |
| Round 7 | 6 | 9 | 67% | Phase NN |
| Season | 45 | 63 | 71% |
Model Notes
Hawthorn vs Gold Coast was the most embarrassing miss. Gold Coast rated 56% on the road at UTAS — basically a coin flip — and Hawthorn won by 49. The model saw this as a genuine contest; the result was anything but. Hawthorn’s rolling numbers still look moderate but they keep winning by large margins when they win.
Geelong (79%) losing to Port by 30 is a pattern. Port Adelaide at Adelaide Oval have now beaten two high-confidence opponents this season. The model weights travel and rolling averages heavily but Adelaide Oval’s home-ground effect may be stronger than the features capture.
The low-confidence calls (54%, 56%) going wrong is expected. In that range you’re essentially flipping a coin. Western Bulldogs at 54% losing by 66 is a bad outcome but not a model failure — the model never had conviction here.
GWS won by 7 as a 64% favourite. Right call, uncomfortable margin. They’ve now won three straight games by 1, 7, and 7 points in situations where the model had them as clear favourites. The model is right about the direction but consistently overstates their dominance.
Leave a Reply