Jack Gun Cyril Stun
Club Legend
- Oct 5, 2012
- 1,237
- 1,504
- AFL Club
- Hawthorn
Squiggle can’t appreciate that both games were effectively dusted at half time.
Especially Eagles v Demons.
Especially Eagles v Demons.
Certified Legendary Thread The Squiggle is back in 2023 (and other analytics)
- Thread starter Final Siren
- Start date
- Tagged users None
iluvparis
Import Whisperer
- Apr 1, 2005
- 50,055
- 40,651
- AFL Club
- Carlton
- Other Teams
- Calgary Flames, Man Utd
- Aug 18, 2009
- 4,229
- 17,495
- AFL Club
- Richmond
- Thread starter
- #954
I'm mostly working on a proper simulator at the moment, i.e. this thing:
Very early days, but it's lots of fun to play with!
- Apr 21, 2014
- 3,846
- 6,687
- AFL Club
- Fremantle
- Other Teams
- Vancouver Canucks
Final Siren looking at the groupings I think there is another point we can raise beyond the even-ness of the last few years.
I think a few years ago you did an analysis of attacking vs defending teams, in terms of their squiggle ratings, and the relative premiership numbers won by both.
Barring a huge flood of goals by either side (could happen given what we saw in the prelims), this will be the 3rd year in a row that a team with a Defence > Offence value will have won.
Are we starting to see a change in the long term trend, or is this simply a result of the competition overall becoming more defensive (eyeballing would suggest end of season squiggle distributions are becoming lower and flatter, and we all know scoring is down and the 'state of the game' narrative exists). Given the relative positions of the teams it may be that only 2017 ends up being the true defence > attack outlier, at least in terms of a comparison to the runner up.
- Aug 18, 2009
- 4,229
- 17,495
- AFL Club
- Richmond
- Thread starter
- #956
Correct, because when I investigated, junktime goals turned out to be significant indicators. Obviously they don't have much bearing on the match outcome, but better sides really do tend to keep scoring more in junk time, while weaker sides let their opposition back in.
- Aug 18, 2009
- 4,229
- 17,495
- AFL Club
- Richmond
- Thread starter
- #957
Good question! It's very much in the realm of speculation, though, because there are so few data points. It did look suspicious when so many attack-centric teams defeated defence-centric teams, especially during the Ross Lyon era. But that scenario only comes around every few years, and in 2017, as you say, it went the opposite way.
Across 2017-18 there have only been two high-scoring teams, Adelaide and Melbourne, and both came to surprisingly brutal ends. Is that because that defence is king nowadays, or is it more about the specifics of those teams on those days in those two games? I dunno. Answering those questions requires a more fine-grained analysis than I can provide, so I can't really conclude anything with confidence.
DELIBERATE!!!!!!!!!!!!!!!
I actually want him to bring out a new book!! Maybe more stories in the lexicon or Jennifer Nation worlds?
Jack Gun Cyril Stun
Club Legend
- Oct 5, 2012
- 1,237
- 1,504
- AFL Club
- Hawthorn
Referring to junk time scores. Squiggle only recognises final scores. Final siren gives brief explanation above.
If you halved demons final score which probably still over rates their performance, WCE would shoot right and up on the squiggle.
iluvparis
Import Whisperer
- Apr 1, 2005
- 50,055
- 40,651
- AFL Club
- Carlton
- Other Teams
- Calgary Flames, Man Utd
Thats why I posted ???
Squiggle does appreciate this - its just that humans massively over react to it in the wrong direction.
West Coast +4 v Collingwood.
Darn it. Missed the margin by a point.
Ratings going into next year
1. West Coast 26.4
2. Collingwood 22.7
3. Melbourne 19.4
4. Richmond 12.4
5. Geelong 11.7
6. GWS 9.9
7. Adelaide 8.5
8. Essendon 7.2
9. Hawthorn 2.9
10. North Melbourne 2.5
11. Western Bulldogs -1.9
12. Brisbane -2.2
13. Port Adelaide -2.3
14. Sydney -2.4
15. St Kilda -8.0
16. Fremantle -11.3
17. Gold Coast -16.2
18. Carlton -18.7
There's a skew of about +3.3 per team, because the last nine matches have been between high-rated teams so the average rating is above zero. This doesn't functionally affect the predictions, and will be mostly gone halfway through next season.
Darn it. Missed the margin by a point.
Ratings going into next year
1. West Coast 26.4
2. Collingwood 22.7
3. Melbourne 19.4
4. Richmond 12.4
5. Geelong 11.7
6. GWS 9.9
7. Adelaide 8.5
8. Essendon 7.2
9. Hawthorn 2.9
10. North Melbourne 2.5
11. Western Bulldogs -1.9
12. Brisbane -2.2
13. Port Adelaide -2.3
14. Sydney -2.4
15. St Kilda -8.0
16. Fremantle -11.3
17. Gold Coast -16.2
18. Carlton -18.7
There's a skew of about +3.3 per team, because the last nine matches have been between high-rated teams so the average rating is above zero. This doesn't functionally affect the predictions, and will be mostly gone halfway through next season.
- Aug 18, 2009
- 4,229
- 17,495
- AFL Club
- Richmond
- Thread starter
- #963
2018 Premiers: West Coast Eagles
And here is their journey highlighted:
It was a very even year, when any number of teams might have seized the cup with a dash of form in the finals. The Eagles managed to hang around in striking distance during the season despite injuries, then delivered when it counted.
With +5 scoring shots and a "home ground disadvantage," Squiggle thinks the Eagles deserved to win the Grand Final.
It is the least dominant season from a premier in the last 20 years, since West Coast never ranked #1 on Squiggle metrics and was only briefly #2 after Rounds 9-10. This can also be seen in how five teams finished the season with a percentage of 120%: West Coast's was 121.4%.
So it's a flag that's come from seizing opportunity, rather than outright dominance.
2019
It's never too early to think about next year! Before adjusting for the fixture, trades, delistings, free agent moves, injuries, rehabilitation, and pre-season form, this is Squiggle's form rating leading into 2019:
- Richmond
- Greater Western Sydney
- Melbourne
- Collingwood
- West Coast
- Geelong
- Essendon
- Hawthorn
- Port Adelaide
- Adelaide
- North Melbourne
- Sydney
- St Kilda
- Western Bulldogs
- Brisbane Lions
- Fremantle
- Gold Coast
- Carlton
Squiggle had a good tipping year, especially with the new algorithm, SQUIGGLE4, which took over mid-season:
- SQUIGGLE4: 151 tips (73.0%), 26.37 MAE
- SQUIGGLE2: 142 tips (68.6%), 26.69 MAE
- ISTATE-91:12: 142 tips (68.6%), 27.21 MAE
2018 Premiers: West Coast Eagles
And here is their journey highlighted:
It was a very even year, when any number of teams might have seized the cup with a dash of form in the finals. The Eagles managed to hang around in striking distance during the season despite injuries, then delivered when it counted.
With +5 scoring shots and a "home ground disadvantage," Squiggle thinks the Eagles deserved to win the Grand Final.
It is the least dominant season from a premier in the last 20 years, since West Coast never ranked #1 on Squiggle metrics and was only briefly #2 after Rounds 9-10. This can also be seen in how five teams finished the season with a percentage of 120%: West Coast's was 121.4%.
So it's a flag that's come from seizing opportunity, rather than outright dominance.
2019
It's never too early to think about next year! Before adjusting for the fixture, trades, delistings, free agent moves, injuries, rehabilitation, and pre-season form, this is Squiggle's form rating leading into 2019:
Algorithms
- Richmond
- Greater Western Sydney
- Melbourne
- Collingwood
- West Coast
- Geelong
- Essendon
- Hawthorn
- Port Adelaide
- Adelaide
- North Melbourne
- Sydney
- St Kilda
- Western Bulldogs
- Brisbane Lions
- Fremantle
- Gold Coast
- Carlton
Squiggle had a good tipping year, especially with the new algorithm, SQUIGGLE4, which took over mid-season:
Combined (i.e. using tips from SQUIGGLE2 for Rounds 1-12 and SQUIGGLE4 thereafter), Squiggle scored 146 tips (70.5%), which placed it third on behind AFL Live Ladders and Massey Ratings on the leaderboard of public computer models, who both recorded 147 tips. The Age's expert tips winner also scored 147.
- SQUIGGLE4: 151 tips (73.0%), 26.37 MAE
- SQUIGGLE2: 142 tips (68.6%), 26.69 MAE
- ISTATE-91:12: 142 tips (68.6%), 27.21 MAE
How many tips for just picking favorites?
- Aug 18, 2009
- 4,229
- 17,495
- AFL Club
- Richmond
- Thread starter
- #965
147 if you averaged all the bookies’ closing odds, ie placed your tip at the last second. Fewer if you tip earlier, since you miss last-minute team changes.
Also 147 if you averaged all the models each game (called “Aggregate” on the leaderboard).
Both of these seem to exhibit a “wisdom of crowds” effect, eg Aggregate got 147 even though the models it’s averaging got 138-147!
Up yours Squiggle
Eagles are premiers
Eagles are premiers
But the Tigers are still ahead on the REAL competition - the mighty Squiggle!Up yours Squiggle
Eagles are premiers
Do you think at some stage the simulation might allow us to play around with match ups say the 2017 tigers v 2013 Hawks?
Aramis
Premiership Player
- Jul 31, 2016
- 4,324
- 9,074
- AFL Club
- Hawthorn
What a fantastic idea, would love to see this implemented. Might be an absolute nightmare to produce though.
tazaa
Brownlow Medallist
Yes!!!!2018 Premiers: West Coast Eagles
And here is their journey highlighted:
It was a very even year, when any number of teams might have seized the cup with a dash of form in the finals. The Eagles managed to hang around in striking distance during the season despite injuries, then delivered when it counted.
With +5 scoring shots and a "home ground disadvantage," Squiggle thinks the Eagles deserved to win the Grand Final.
It is the least dominant season from a premier in the last 20 years, since West Coast never ranked #1 on Squiggle metrics and was only briefly #2 after Rounds 9-10. This can also be seen in how five teams finished the season with a percentage of 120%: West Coast's was 121.4%.
So it's a flag that's come from seizing opportunity, rather than outright dominance.
2019
It's never too early to think about next year! Before adjusting for the fixture, trades, delistings, free agent moves, injuries, rehabilitation, and pre-season form, this is Squiggle's form rating leading into 2019:
Algorithms
- Richmond
- Greater Western Sydney
- Melbourne
- Collingwood
- West Coast
- Geelong
- Essendon
- Hawthorn
- Port Adelaide
- Adelaide
- North Melbourne
- Sydney
- St Kilda
- Western Bulldogs
- Brisbane Lions
- Fremantle
- Gold Coast
- Carlton
Squiggle had a good tipping year, especially with the new algorithm, SQUIGGLE4, which took over mid-season:
Combined (i.e. using tips from SQUIGGLE2 for Rounds 1-12 and SQUIGGLE4 thereafter), Squiggle scored 146 tips (70.5%), which placed it third on behind AFL Live Ladders and Massey Ratings on the leaderboard of public computer models, who both recorded 147 tips. The Age's expert tips winner also scored 147.
- SQUIGGLE4: 151 tips (73.0%), 26.37 MAE
- SQUIGGLE2: 142 tips (68.6%), 26.69 MAE
- ISTATE-91:12: 142 tips (68.6%), 27.21 MAE
We are no longer the weakest premier
Thank you West Coast and Collingwood!
Lost in the prelim but don’t care because we’re the best on the squiggle
iluvparis
Import Whisperer
- Apr 1, 2005
- 50,055
- 40,651
- AFL Club
- Carlton
- Other Teams
- Calgary Flames, Man Utd
2018 Premiers: West Coast Eagles
And here is their journey highlighted:
It was a very even year, when any number of teams might have seized the cup with a dash of form in the finals. The Eagles managed to hang around in striking distance during the season despite injuries, then delivered when it counted.
With +5 scoring shots and a "home ground disadvantage," Squiggle thinks the Eagles deserved to win the Grand Final.
It is the least dominant season from a premier in the last 20 years, since West Coast never ranked #1 on Squiggle metrics and was only briefly #2 after Rounds 9-10. This can also be seen in how five teams finished the season with a percentage of 120%: West Coast's was 121.4%.
So it's a flag that's come from seizing opportunity, rather than outright dominance.
2019
It's never too early to think about next year! Before adjusting for the fixture, trades, delistings, free agent moves, injuries, rehabilitation, and pre-season form, this is Squiggle's form rating leading into 2019:
Algorithms
- Richmond
- Greater Western Sydney
- Melbourne
- Collingwood
- West Coast
- Geelong
- Essendon
- Hawthorn
- Port Adelaide
- Adelaide
- North Melbourne
- Sydney
- St Kilda
- Western Bulldogs
- Brisbane Lions
- Fremantle
- Gold Coast
- Carlton
Squiggle had a good tipping year, especially with the new algorithm, SQUIGGLE4, which took over mid-season:
Combined (i.e. using tips from SQUIGGLE2 for Rounds 1-12 and SQUIGGLE4 thereafter), Squiggle scored 146 tips (70.5%), which placed it third on behind AFL Live Ladders and Massey Ratings on the leaderboard of public computer models, who both recorded 147 tips. The Age's expert tips winner also scored 147.
- SQUIGGLE4: 151 tips (73.0%), 26.37 MAE
- SQUIGGLE2: 142 tips (68.6%), 26.69 MAE
- ISTATE-91:12: 142 tips (68.6%), 27.21 MAE
Starting stacks for 2019?
2018 Premiers: West Coast Eagles
And here is their journey highlighted:
It was a very even year, when any number of teams might have seized the cup with a dash of form in the finals. The Eagles managed to hang around in striking distance during the season despite injuries, then delivered when it counted.
With +5 scoring shots and a "home ground disadvantage," Squiggle thinks the Eagles deserved to win the Grand Final.
It is the least dominant season from a premier in the last 20 years, since West Coast never ranked #1 on Squiggle metrics and was only briefly #2 after Rounds 9-10. This can also be seen in how five teams finished the season with a percentage of 120%: West Coast's was 121.4%.
So it's a flag that's come from seizing opportunity, rather than outright dominance.
2019
It's never too early to think about next year! Before adjusting for the fixture, trades, delistings, free agent moves, injuries, rehabilitation, and pre-season form, this is Squiggle's form rating leading into 2019:
Algorithms
- Richmond
- Greater Western Sydney
- Melbourne
- Collingwood
- West Coast
- Geelong
- Essendon
- Hawthorn
- Port Adelaide
- Adelaide
- North Melbourne
- Sydney
- St Kilda
- Western Bulldogs
- Brisbane Lions
- Fremantle
- Gold Coast
- Carlton
Squiggle had a good tipping year, especially with the new algorithm, SQUIGGLE4, which took over mid-season:
Combined (i.e. using tips from SQUIGGLE2 for Rounds 1-12 and SQUIGGLE4 thereafter), Squiggle scored 146 tips (70.5%), which placed it third on behind AFL Live Ladders and Massey Ratings on the leaderboard of public computer models, who both recorded 147 tips. The Age's expert tips winner also scored 147.
- SQUIGGLE4: 151 tips (73.0%), 26.37 MAE
- SQUIGGLE2: 142 tips (68.6%), 26.69 MAE
- ISTATE-91:12: 142 tips (68.6%), 27.21 MAE
A possible quirky little tidbit but Squiggle forecast the correct final scores, just allocated them to the wrong teams.
I think you'll find it's harder to win a flag when the competition is so even.
The weakest premier has to be Geelong 2007 doesn't it?
