Finals predictions - mid-term form, short-term form and winning streaks.

[Hobbes]

Just a quick statistical dive, taking data from teams finishing 1-4 (who make up well over 90% of grand finalists) based on: W/L from the last ten matches, W/L from the last five matches, and winning/losing streaks heading into the finals. I'm just noting the point of exit rather than examining each final - that would be more work.

Mid-term form (last 10 matches)

10 wins: 2 Premiers, one Runner-up, one Prelim exit (Adelaide in 2005)
9 wins: 5 Premiers, three Runners-up, 9 Prelims, one straight sets exit (Brisbane 2019)
8 wins: 8 Runners-Up, 10 Prelims, 7 Premiers, 4 straight sets.
7.5 wins: 1 Prelim, 1 Runner-up
7 wins: 5 Runners-up, 11 Prelim, 4 Premier, 5 straight sets.
6.5 wins: 2 Runners-up, 1 Premier
6 wins: 1 runner-up, 8 Prelims, 4 Premiers, 2 straight sets.
5 wins: 2 runner-up, 2 prelim

Short-term form (last 5 matches)

5 wins: 4 Runners-up, 13 prelims, 4 premiers, 1 straight sets (Hawthorn 2018)
4 wins: 9 Runners-up, 16 prelims, 15 premiers, 8 straight sets
3 wins: 7 Runners-up, 6 prelims, 2 premiers, 3 straight sets
2.5 wins: 1 Runner-up.
2 wins: 2 Runners-up, 7 prelims, 2 premiers

Winning streaks (Current at end of season)

13 wins: 2 premiers, 1 runner-up. (Brisbane 2001, Geelong 2008/2022)
10 wins: 1 prelim (Adelaide 2005)
9 wins: 1 prelim, 1 premier (Richmond 2019)
8 wins:1 prelim (Port Adelaide 2003)
7 wins: 1 runner-up, 4 prelims.
6 wins: 1 runner-up, 5 prelims, 1 premier (Richmond 2020), 1 straight sets (Hawthorn 2018)
5 wins: 1 runner-up, 1 prelim
4 wins: 3 runners-up, 2 prelims, 3 premiers, 2 straight sets
3 wins: 4 runners-up, 5 prelims, 1 premier, 3 straight sets
2 wins: 1 runner-up, 4 prelims, 5 premiers, 2 straight sets
1 win: 3 runners-up, 12 prelims, 7 premiers, 2 straight sets
1 loss: 6 runners-up, 5 prelims, 2 premiers, 2 straight sets
2 losses: 2 runners-up, 1 prelim, 1 premier

Conclusions aren't obvious, but I do notice that teams that finished with 1-2 wins won 12 premierships (out of 25) but only lost four grand finals.

Grand Final head-to-heads (I can only take into account matchups where both of the teams were top-4).

Short term form win: 2022, 2020, 2017, 2015, 2010, 2005, 2004, 2002, 2001
Short term form lose: 2023, 2012, 2008, 2003, 2000
Mid-term form win: 2022, 2017, 2015, 2013, 2010, 2007, 2005, 2004, 2002, 2001, 2000
Mid-term form lose: 2023, 2018, 2012, 2011, 2009, 2008
Winning streak win: 2022, 2020, 2017, 2015, 2014, 2013, 2011, 2005, 2004, 2001
Winning streak lose: 2023, 2018, 2012, 2008, 2007, 2006, 2003, 2000

Short term+Mid term win: 2022, 2017, 2015, 2010, 2005, 2004, 2002
Short term+Mid term lose: 2023, 2012, 2008
Short term+streak win: 2022, 2020, 2017, 2015, 2005, 2004
Short term+streak lose: 2023, 2012, 2008, 2003
Mid term+streak win: 2022, 2017, 2015, 2013, 2005, 2004
Mid term+streak lose: 2023, 2018, 2012, 2008
All three win: 2022, 2017, 2015, 2005, 2004
All three lose: 2023, 2012, 2008

Short term form: 9/14: 64%
Mid term form: 11/17: 64%
Winning streak: 10/18: 56%
Short+mid: 7/10: 70%
Short+streak: 6/10: 60%
Mid+streak: 6/10: 60%
Short+Mid+Streak: 5/8: 62%

Interesting that all six are positive predictors of notoriously tricky Grand Finals.

Compare with my conclusions from this ignored thread https://www.bigfooty.com/forum/threads/tipping-finals.1397328/

Minor round win/loss: 57%
Percentage: 40%
Home state: 64%
Last head-to-head: 40%

Win/loss is a mild positive predictor, home state is a better one, but percentage and last head-to-head are actually negative predictors of a Grand Final winner.

 

[Fadge]

What?

 

[Corn Cobbers]

What?

Haven't felt this out of my league since foolishly doing that statistics unit at uni

 

[Corn Cobbers]

All I remember is "co-efficient of skewness"

 

[Hobbes]

What?

It’s basically “how much does formline running into the finals predict finals performance?”. Sorry, the presentation is a little dry.

I looked at teams finishing in the top four after the minor round from 2000-2024. For each team, I noted their results (W/L only) from their last five games (short-term), their last ten games (mid-term) and their streak of wins/losses running into the finals (which could be as much as 13 wins or 2 losses - no team on three consecutive losses has finished top four in this period). Then I noted how each of these teams had performed - Premier, Runner-up, Prelim loser or bundled out in the semi-finals.

As an addendum, I looked at grand-final matchups and whether you could predict the results based on better 5-game form, better 10-game form, or longer winning streak than your opponent leading into finals. All three measures are positive predictors of grand final success. This is interesting because (as per my earlier thread, linked above) some things you would think count for something are actually negative indicators in grand finals specifically. Two examples of things that make you less likely to win a grand final are: having a better percentage than your opponent, and winning your last head-t0-head matchup with your opponent.