Just a quick statistical analysis of what predicts a finals result. I've used every finals game from 2000 onward, when the current finals system was adopted, to 2024.
Overall stats:
(I have not counted variables I couldn't split, like being tied for wins, either both home or both away, or if their last head-to-head result was a draw.)
Tipping by HOA Win/Loss record: 138/61
Tipping by Percentage: 138/67
Tipping by Home/Away: 103/48 (I'm defining this by state, so a Victorian team playing a non-Victorian team at the MCG is deemed to have home-ground advantage, but Collingwood are deemed not to have home-ground advantage against Geelong.)
Tipping by last head-to-head result: 116/108
I've also split these by different types of final: Elimination, Qualifying, Semi, Preliminary and Grand.
Elimination finals:
Win/loss: 24/17
Percentage: 28/22
Home: 23/11
Head-to-head: 25/24
Qualifying finals
Win/loss: 31/12
Percentage: 31/19
Home: 22/9
Head-to-Head: 29/21
Semi-finals
Win/Loss: 37/12
Percentage: 33/17
Home: 22/9
Head-to-Head: 26/24
Preliminary Final
Win/Loss: 33/10
Percentage: 36/14
Home: 21/11
Head-to-Head: 26/24
Grand Finals
Win/Loss: 13/10
Percentage: 10/15
Home: 9/5
Head-to-Head: 10/15
What I've learned from this is:
Win/loss is better at predicting finals than percentage. (That wasn't obvious to me). But W/L, percentage and home ground are better than 2/3.
Head-to-head is barely better than throwing a coin.
Grand finals are weird. Win/loss is a little better than 50%, home ground advantage (usually, a Victorian team playing a non-Victorian team) is quite good, but percentage and head-to-head are negative indicators.
Overall stats:
(I have not counted variables I couldn't split, like being tied for wins, either both home or both away, or if their last head-to-head result was a draw.)
Tipping by HOA Win/Loss record: 138/61
Tipping by Percentage: 138/67
Tipping by Home/Away: 103/48 (I'm defining this by state, so a Victorian team playing a non-Victorian team at the MCG is deemed to have home-ground advantage, but Collingwood are deemed not to have home-ground advantage against Geelong.)
Tipping by last head-to-head result: 116/108
I've also split these by different types of final: Elimination, Qualifying, Semi, Preliminary and Grand.
Elimination finals:
Win/loss: 24/17
Percentage: 28/22
Home: 23/11
Head-to-head: 25/24
Qualifying finals
Win/loss: 31/12
Percentage: 31/19
Home: 22/9
Head-to-Head: 29/21
Semi-finals
Win/Loss: 37/12
Percentage: 33/17
Home: 22/9
Head-to-Head: 26/24
Preliminary Final
Win/Loss: 33/10
Percentage: 36/14
Home: 21/11
Head-to-Head: 26/24
Grand Finals
Win/Loss: 13/10
Percentage: 10/15
Home: 9/5
Head-to-Head: 10/15
What I've learned from this is:
Win/loss is better at predicting finals than percentage. (That wasn't obvious to me). But W/L, percentage and home ground are better than 2/3.
Head-to-head is barely better than throwing a coin.
Grand finals are weird. Win/loss is a little better than 50%, home ground advantage (usually, a Victorian team playing a non-Victorian team) is quite good, but percentage and head-to-head are negative indicators.
