Round 10 Predictions

For the third week in a row, my ELO model managed 6 correct tips, with an MAE of 29, bringing our season total to 58 (72%) with an MAE of 31.5.

This upcoming weekend shows us a few close matches that are difficult to call. Our current ELO rankings show us that there is more congestion at the top of the ladder, with a clump of 6 teams forming, with the Bulldogs and North (still!) just outside of that. We again have a decent drop-off after this top 8, suggesting that making the finals from here might be tricky for those bottom teams.


Onto this weeks predictions, and there are a few tricky games to pick. Unfortunately, the only game between our top 6 clump of teams is Adelaide v GWS with our model given the Crows a slight advantage at home. The model also expects tight tussles between StK and Fremantle, essentially tipping a tie (it gives Freo a 50.1% chance of winning) and struggling to seperate Melbourne and PA.

The biggest round of the match, in terms of media coverage, is certainly Sydney v North Melbourne, given that the Kangaroos are still undefeated. Our model is giving the Kangaroos only a 39% chance of extending their start to the season to 10 wins, which would give them the equal 9th best start to a season in AFL history.


Simulating the season gives us a new leader on the top of our predictions - GWS! Our models sees them finish on top in 17% of simulations, just above a group of 4 teams (Geelong, Sydney, North Melbourne and Hawthorn) on 16%. Certainly that log jam at the top is difficult to seperate this far out.


Another interesting pattern is solidifying, where we see a really big split after the top 8. Again, as our ratings suggest, there is a big gulf between those top 8 teams and we don’t give them bottom teams a very big chance of moving up into that top 8. Melbourne (26%) and Port Adelaide (24%) are still the biggest chances.


comments powered by Disqus