Eight from nine. Season at 72.2%.
Results
8/9 correct. The one miss: Melbourne (32%) beat Brisbane by 2 at the MCG. The standout: West Coast were rated 76% home favourites in the derby and lost to Fremantle by 56 — the right team, the completely wrong probability. Hawthorn’s 81% call survived by 3 points; Adelaide’s 64% call survived by 1.
Technical
| Match | Predicted | Prob | Result | Margin |
|---|---|---|---|---|
| Carlton vs Collingwood | Collingwood | 56% | ✓ | −5 |
| Geelong vs W. Bulldogs | Geelong | 64% | ✓ | +75 |
| Sydney vs GWS | Sydney | 64% | ✓ | +41 |
| Gold Coast vs Essendon | Gold Coast | 80% | ✓ | +9 |
| Hawthorn vs Port Adelaide | Hawthorn | 81% | ✓ | +3 |
| Adelaide vs St Kilda | Adelaide | 64% | ✓ | +1 |
| North Melbourne vs Richmond | North Melbourne | 78% | ✓ | +75 |
| Melbourne vs Brisbane Lions | Brisbane Lions | 68% | ✗ | −2 |
| West Coast vs Fremantle | Fremantle | 88% | ✓ | −56 |
Avg confidence: 71.2%.
Season: 39/54 (72%)
Rounds 1–5 used the logistic model; Round 6 onwards uses the Phase NN model.
| 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 |
| Season | 39 | 54 | 72% |
Model Notes
West Coast are breaking the model. They were rated 76% favourites at home in the derby and lost by 56. Their recent results — losses of 128, 46, and 56 — are so extreme that the model’s home-ground and rolling-average features can’t compensate. Every West Coast home game is a model liability until their numbers update.
Geelong are underrated. Three wins by 8, 46, and 75 points across Rounds 3, 5, and 6. The model has been calling their games at 49–62% and they’re winning comfortably. Their 2025 data is dragging the number down. Add a mental nudge upward on Geelong until the rolling window clears.
Melbourne remain unreadable. They’ve now beaten Gold Coast, Carlton, and Brisbane (all upsets or close calls) while losing badly to West Coast and Essendon. No coherent signal. Wide error bars on any Melbourne prediction — the model is effectively guessing in this range.
Hawthorn’s wins keep being close. 96% confidence, 3-point win. The model is right about the direction but something about their actual performance suggests the margin of superiority is being overstated. Five wins from six is elite, but the close finishes (1-point vs Geelong R4, 3-point vs Port R6) suggest they’re performing closer to 70–75% than 96%.
Leave a Reply