Certified Legendary Thread The Squiggle is back in 2023 (and other analytics)

Found my ladder prediction after I predicted every match back in March, and chucked it into the 'Rate My Ladder' feature on Squiggle (nice addition by the way). Fair to say I was a little off

734164
 
Sep 16, 2016
1,888
3,290
I have no idea
AFL Club
Western Bulldogs
Other Teams
Quite a few

Final Siren

Mr Squiggle
Aug 18, 2009
4,229
17,495
AFL Club
Richmond
It’s based purely on number of games played there over a certain period (don’t know what that period is exactly). Would come pretty damn close to zero when two tenants meet I’d say. Last night was +1.7pts to Collingwood and Essendon play most home games at Etihad.
Yep, this is correct. Squiggle looks at 3 previous years plus the current year.

Different models do HGA in different ways. It's hard to know who's right, since they usually agree within a handful of points, which is too small to test rigorously.

So we can be confident that an HGA estimate for a particular match is about right - within a goal, for most matches - but we can't really be sure whether it should be a couple of points higher or lower.
 

Final Siren

Mr Squiggle
Aug 18, 2009
4,229
17,495
AFL Club
Richmond
You both beat Damien Barrett!

https://squiggle.com.au/rate-my-ladder/?experts=1

Predicting ladders is hard, but even so, Barrett's ladder is terrible. It's the worst one I've seen on the internet. He managed to get only 4 finalists, had Fremantle as his big improver, and tipped Adelaide to take out the minor premiership, along with the more common misses like having Melbourne far too high and Brisbane far too low.

If you closed your eyes and randomly placed teams into ladder positions, you would do better than Barrett 10% of the time.
 

hk89

Norm Smith Medallist
Jul 12, 2009
8,503
10,459
Melbourne
AFL Club
Hawthorn
Yep, this is correct. Squiggle looks at 3 previous years plus the current year.

Different models do HGA in different ways. It's hard to know who's right, since they usually agree within a handful of points, which is too small to test rigorously.

So we can be confident that an HGA estimate for a particular match is about right - within a goal, for most matches - but we can't really be sure whether it should be a couple of points higher or lower.

Have you done any modelling on home crowd sizes vs pure HGA? I often wonder how much of the HGA is down to local knowledge of the ground vs the impact of the crowd support on the local team (and the umpires). Probably not that useful for predictive modelling given you don't the crowd ahead of time, but might be interesting to know if there is a pure crowd size effect (and to a degree crowd sizes are probably somewhat predictable, so might not have zero predictive value, although probably not enough to be practically useful).
 

Final Siren

Mr Squiggle
Aug 18, 2009
4,229
17,495
AFL Club
Richmond
Who is the best squiggle team to not make finals? Hawthorn surely up there this year.
GWS game helped a bit
Mmm, I may tweak this in retrospect because I think the lower scoring has created a pretty significant distortion.

But anyway, Geelong 1993 are the best team to not make finals -- finished Home & Away as the top-rated team in an absolute car crash season of footy.
 

Final Siren

Mr Squiggle
Aug 18, 2009
4,229
17,495
AFL Club
Richmond
Have you done any modelling on home crowd sizes vs pure HGA? I often wonder how much of the HGA is down to local knowledge of the ground vs the impact of the crowd support on the local team (and the umpires). Probably not that useful for predictive modelling given you don't the crowd ahead of time, but might be interesting to know if there is a pure crowd size effect (and to a degree crowd sizes are probably somewhat predictable, so might not have zero predictive value, although probably not enough to be practically useful).
I haven't tried it since, as you say, it's not very helpful for prediction as you don't know crowd sizes in advance. But yes, I do believe from what I've read that crowd size (& noise/passion) is likely to be a very significant factor - even more important than things like travel & ground familiarity.
 

Final Siren

Mr Squiggle
Aug 18, 2009
4,229
17,495
AFL Club
Richmond

Pessimistic

Cancelled
30k Posts 10k Posts HBF's Milk Crate - 70k Posts TheBrownDog
Sep 13, 2000
86,852
42,951
Melbourne cricket ground. Australia
AFL Club
Hawthorn
Other Teams
Horks
If you call hawthorn and geelong outliers (so few flags won on the defensive flank) a collingwood richmond grand final sounds likely.

But heres a question. Who deserves top spot ie the home strip?
 

Pessimistic

Cancelled
30k Posts 10k Posts HBF's Milk Crate - 70k Posts TheBrownDog
Sep 13, 2000
86,852
42,951
Melbourne cricket ground. Australia
AFL Club
Hawthorn
Other Teams
Horks
The Squiggle's forecast has already eliminated Essendon!!

Has a finalist been so poor in squiggle?

Also compare 2016 and 2019. 2016 was a close finals because so many team in premiership zone. 2019 close because they arent in the zone at all
 

Corkintheocean

Premiership Player
Aug 15, 2009
3,558
5,460
AFL Club
West Coast
For curiousities sake, is the model able to chart West Coast games not involving Naitanui vs West Coast with Naitanui over say the last three seasons?
 
Back