AI Game Review
Our model correctly predicted Collingwood to win at 62% probability. The margin model missed here — predicting 32.2 but the actual margin was 56 points. The game's 174 points came in 16 points higher than the predicted 158. Collingwood led 29–53 at the break and pulled away in the second half to win by 56. The model went 2/3 on this match. The 1-39 margin band call landed.
Model vs actual outcomes • Post-match analysis
Quarter-by-Quarter Win Probability
AI Win Probability
Carlton
38%
Collingwood
62%
AI Match Overview
Collingwood hold the advantage at 62% win probability, though Carlton are far from out of this at 38%. The model sees Collingwood ahead on 5 of 7 key factors including ELO Difference, Midfield ELO and Recent Win Rate. Collingwood carry a 482-point ELO rating advantage (1841 vs 1359). Recent form favours Collingwood with 5 wins from their last 5 compared to 2 for Carlton. The margin model predicts Collingwood by 32.2 points with a combined total of 158.
Generated from model features • Pre-kick-off analysis
Edge Analysis
Each market is predicted by an independent model — H2H, margin, and totals may occasionally disagree.
H2H Recommendation
Collingwood to Win @1.22
Winner ✓
Edge
-20.4%
Line / Spread
Collingwood +27.5 @1.91
Winner ✓
Edge
-20.4%
Total Points
Under 161.5 @1.91
Lost ✗
Edge
+2.6%
Form & History
| Team | Last 5 | Avg Pts |
|---|---|---|
Carlton | WWLLL | 75.8 |
Collingwood | WWWWW | 96.6 |
Avg Conceded
85.2
Carlton
64.6
Collingwood
Avg Margin
-9.4
Carlton
32.0
Collingwood
Disposals
361.2
Carlton
347.8
Collingwood
Inside 50s
50.0
Carlton
50.0
Collingwood
📊Team ELO Ratings
🏈Positional Matchups
Player ELO aggregated by position group — higher = stronger unit
📈Recent Form (Last 5)
🔑Key Prediction Factors
What the model weighted most in this prediction
Model Confidence
62%
Collingwood predicted to win by 32 points
Predicted total: 158 · Line: -32.2
Player Work Effort
Per-minute effort vs effectiveness (vs personal average)Team Effort
-0.16
Team Effectiveness
-0.01
Effort = pressure acts + tackles + contested possessions per minute on field, z-scored vs career avg. Effectiveness = disposal efficiency + fantasy/min + score involvements − errors, z-scored vs career avg.
Goal Scorer Predictions
AI-powered goal scorer predictions and player prop markets — built on our 6-model player stats engine.