Certified Legendary Thread Squiggle 2017

[WhiskeyTangoFoxtrot]

Has them all at the right end of the ladder though!

Correct. However, I'm wondering what the impact is on the rest of the results if Essendon is under-rated (much like Freo was over-rated early in 2016). For example, up until Round 10, the Squiggle gives Essendon a win % of at most 35% in any game (with most in the 20-30% range). That just seems a little low to me.

I guess the flip side of that is that any side that loses to Essendon will take a big hit to their Squiggle rating as they will lose roughly 0.75 of a win currently predicted for them.

 

[Komrade_Tiger]

Squiggle doesnt like Essendons chances, i just missed catching it at 96% as well!

 

[igon value]

I still think they're a little under-rated. The Squiggle currently has them finishing equal second last with 7 wins.

I think, despite yesterday's result, Essendon will finish higher than Carlton, Gold Coast, Fremantle, Brisbane and, if they keep playing the way they are currently, Hawthorn.

At this point "equal second last" is meaningless. Essendon are in a group that could finish ~13-17th, which I think is fair.

Squiggle is probably overpricing Hawthorn, true. But two wins split Hawks and Essendon - I'd give Essendon one or two more wins and Hawthorn one less, but otherwise it doesn't seem that far off.

 

[richoatthedisco]

Seems harsh but I guess Tigers have a history and have yet to leave Melbourne. Hope we play you in a grand final

Me too. Must be 20 years since you've beaten us at the G.

 

[Final Siren]

Final Siren, given the large personnel change at Essendon from 2016 to 2017, do you think the Squiggle currently underrates them considerably? If so, how many weeks do you think it will take for that effect to fall away?

I don't think Essendon are especially underrated. There are often teams that rise or fall spectacularly from one year to the next, for whatever reason, so the possibility of that kind of thing is already baked into the model. I don't think Essendon's circumstances are so unique that they break the upper bound on that, even though the whole suspension thing is unprecedented.

The thing is there's always something you can point at as evidence that this time is different. Essendon regaining players from suspension is different, sure, but we've also had two expansion sides who started from zero and loaded up on young draft picks, as well as plenty of sides that weren't very motivated, for whatever reason, to perform well in the back end of the season prior. Those are part of the data set the model uses to figure out how quickly team form can change.

Most times, this time isn't different, except in the details. That's the whole point of modelling, to cut through the superficial differences and dig out the underlying forces.

The Bombers might turn it on this season and rise a lot, but I don't think they're being unfairly treated at the moment any more than, say, GWS last year, or the Bulldogs the year before that.

If they really are a completely transformed team, squiggle will probably underrate them for quite a while, because genuinely transformed teams are rare, and squiggle will be cynical about whether Essendon are one of them.

 

[WhiskeyTangoFoxtrot]

I don't think Essendon are especially underrated. There are often teams that rise or fall spectacularly from one year to the next, for whatever reason, so the possibility of that kind of thing is already baked into the model. I don't think Essendon's circumstances are so unique that they break the upper bound on that, even though the whole suspension thing is unprecedented.

The thing is there's always something you can point at as evidence that this time is different. Essendon regaining players from suspension is different, sure, but we've also had two expansion sides who started from zero and loaded up on young draft picks, as well as plenty of sides that weren't very motivated, for whatever reason, to perform well in the back end of the season prior. Those are part of the data set the model uses to figure out how quickly team form can change.

Most times, this time isn't different, except in the details. That's the whole point of modelling, to cut through the superficial differences and dig out the underlying forces.

The Bombers might turn it on this season and rise a lot, but I don't think they're being unfairly treated at the moment any more than, say, GWS last year, or the Bulldogs the year before that.

If they really are a completely transformed team, squiggle will probably underrate them for quite a while, because genuinely transformed teams are rare, and squiggle will be cynical about whether Essendon are one of them.

Cheers for the feedback and I can certainly understand what you're saying. I guess time will tell how accurately the modelling is in respect to the Bombers.

Hopefully it holds true for this week at least and the Crows give them a spanking! Looking at the expected result (132-60), the Crows will need to perform exceptionally well both offensively and defensively just to maintain their current Squiggle position.

 

[beert]

OT, but if anyone deserves a free plug, it's Final Siren.

I'm just about to jump on a plane, and was looking for something to read, so picked up one of Final Siren's (Max Barry's) books. (Lexicon). Started it last night - pretty good. A bit like Matthew O'Reilly, but much better characterisation. Only fault is a disappointing lack of squiggles. Definitely recommended - I might need to buy another one for the trip home, this one is a page-turner.

 

[Jars458]

Me too. Must be 20 years since you've beaten us at the G.

2008. Only played there twice since then. I think we will manage

 

[Final Siren]

Round 3, 2017

Animated!

So... Hawthorn. This isn't an easy conversation. You've been there for me since I started posting squiggle stuff on BigFooty. You camped out among the premiership cups for months on end. You were a pleasure to watch and easy to predict.

It's even easy to identify the moment it all went wrong for Hawthorn: Round 6, April 30, 2016, 4:35pm. That was when the Hawks, travelling at 4 wins and 1 loss after a series of narrow escapes, ran into GWS and started watching footballs flying over their heads. They lost by 75 points. It was the beginning of the end because if there was one thing the Hawks made clear during their period of dominance, it was that good teams don't get thrashed.

Thrashings are very indicative. Squiggle pays a lot of attention to them, even more than wins. Very good teams don't receive them, under any circumstances, not even when nothing's going right. They cough up close losses but not thrashings. And they belt bad teams.

So after copping an 86-point belting from the Suns, the Hawks take the big plunge on the squiggle, while Gold Coast leap toward the middle of the pack:

Adelaide and GWS did what was expected, while Sydney had another bad week -- again, not because their performance was so terrible, but because when you can only afford to lose 5 or 6 games for the year to make the top 4, it's not great to accumulate three of them before your first win. They're still rated the third-best team, but they probably need to go at least 16-3 from here, and that's a huge ask.

A good week for the Cats, mainly because Sydney and the Bulldogs' troubles freed up some breathing space in the top 4. Also good for Collingwood, who banked a win that few would have counted on, and stole ground from top-8 competitors in Hawthorn, Melbourne, and North.

Carlton and Essendon take a step to the right, mainly because of a heavily rain-affected game. But Fremantle overtake both after beating the Bulldogs.

The Tigers are now 3-0, but squiggle isn't too excited yet. They're a genuine finals chance, currently predicted to miss out on percentage only, but their results have been only mildly better than expectation. To really move, they need to belt Brisbane next week.

Flagpole! Pretty much business as usual here:

Live squiggle!

Squiggle dials!

 

[Final Siren]

Cheers for the feedback and I can certainly understand what you're saying. I guess time will tell how accurately the modelling is in respect to the Bombers.

Hopefully it holds true for this week at least and the Crows give them a spanking! Looking at the expected result (132-60), the Crows will need to perform exceptionally well both offensively and defensively just to maintain their current Squiggle position.

#firstonflagpoleproblems

 

[Final Siren]

I wonder if you can increase your data points to include players. That way you can take into account personnel changes in the offseason. I think you could even factor in expected output from a draft position plus expect improvement in the first 5 year and expected loss of output after 28. Perhaps an injury or time out of the game could lead to a drop in expected output.

The Arc has just started testing this: https://thearcfooty.com/2017/04/07/fantasy-points-what-are-they-good-for/

He's getting pretty good early results: 70% tipping based on fantasy points alone. That's very promising, I reckon.

The problem, as usual, is that the AFL & Champion Data have most of the useful stats locked up tight, so no-one else can analyze them. This drives the amateur stats guys crazy, because they could do amazing things if the data were available. But it's not.

So if you want to build a player-based model, it has to rely only on the crudest metrics, like disposals.

I hope your modelling is picked up by a news organisation so the can fund you to do this full time, as well as giving you computer resources to crunch more data and people to help gather more data for you. Keep it up bro!

Thanks! I don't want to brag but I just got an email from an actual AFL club saying they're enjoying the squiggle. So that's pretty cool. I mean, wtf an actual AFL club is doing on BigFooty, I don't know. But still. Awesome.

 

[Clems Knee]

Thanks! I don't want to brag but I just got an email from an actual AFL club saying they're enjoying the squiggle. So that's pretty cool. I mean, wtf an actual AFL club is doing on BigFooty, I don't know. But still. Awesome.

Has to be Adelaide. No way my club would have been enjoying the squiggle for the last 18 months.
Seriously though, well done.

 

[NikkiNoo]

Has to be Adelaide. No way my club would have been enjoying the squiggle for the last 18 months.
Seriously though, well done.

The Squiggle has loved Adelaide for a long time and we have loved it back (mostly). We know that the club has in the past closely monitored the BF boards and we meet with the club once a year as well to ask them questions and give (polite) feedback (much to some posters annoyance who think we should tell them we think are bad at things). I wouldn't be surprised if it was my club.

 

[WhiskeyTangoFoxtrot]

Round 3, 2017

Animated!

GWS's squiggle looks like it's doing laps around the block. Bulldogs and Swans are on a rollercoaster, while Hawthorn is just a slippery dip!​

 

[mattsea]

The Arc has just started testing this: https://thearcfooty.com/2017/04/07/fantasy-points-what-are-they-good-for/

He's getting pretty good early results: 70% tipping based on fantasy points alone. That's very promising, I reckon.

Cheers! Yeah I plan to do a lot more work on this when I get the time. As you said, the results with just using Fantasy points are pretty encouraging. There are a lot of stats (eg. cont poss, goal assists) that are publicly available but not incorporated in Fantasy numbers; my next step is to incorporate these are derive some better weights for the stats than the standard ones for Fantasy points. After that, I'll start trying to get into projections based on age, etc. This is a rabbit warren but I think that before next season I should (hopefully) have some useful player-based method for projecting team performance, ideally one that complements my Elo model.

 

[GreyCrow]

OT, but if anyone deserves a free plug, it's Final Siren.

I'm just about to jump on a plane, and was looking for something to read, so picked up one of Final Siren's (Max Barry's) books. (Lexicon). Started it last night - pretty good. A bit like Matthew O'Reilly, but much better characterisation. Only fault is a disappointing lack of squiggles. Definitely recommended - I might need to buy another one for the trip home, this one is a page-turner.

Get your pen out and make your own Squiggles

 

[Billy ray]

is that the largest single game movement (to the positive!) for the suns?

 

[lamaros]

Yeah turns out Essendon isn't under-rated, Hawks are just really really bad.

Curses.

 

[Handsome.B.Wonderful]

The Arc has just started testing this: https://thearcfooty.com/2017/04/07/fantasy-points-what-are-they-good-for/

He's getting pretty good early results: 70% tipping based on fantasy points alone. That's very promising, I reckon.

Read The Arc's post with interest last week and compared it with the @AFLPlayerRating CD selected team quality for each game.

3 wins / 6 losses for Rd 3.

 

[drummond2fevola]

Brisbane really making strides on the squiggle this week.

 

[KissKiss]

Brisbane really making strides on the squiggle this week.

Positive movement is a great change for us

 

[Rotayjay]

Look at that very tight cluster near the middle of the chart. Geelong is at its head. Includes Port, West Coast, Collingwood, Hawthorn, North, Dees and St Kilda - that's eight sides.

 

[vide0n]

Go Suns!

 

[TheMcManusNose]

Really interesting if 17th ends up on 7 wins. I can't recall the last year the competition was that even. Surely pre-GC.

 

[Hobbes]

Really interesting if 17th ends up on 7 wins. I can't recall the last year the competition was that even. Surely pre-GC.

Wins at the bottom of the table

2016: 3,3,4,6
2015: 4,4,4.5,6
2014: 4,4,6,7
2013: 1,1,5,8
2012: 2,2,4,5
2011: 3,3,4,7
2010: 4,6,6,7
2009: 4,5.5,6,7.5
2008: 3,4,6,7
2007: 3,4,5,9.5
2006: 3.5,3.5,7,7
2005: 4.5,5,5,8
2004: 4,4,5,8
2003: 3.5,4,5,7
2002: 3,5.5,7,9
2001: 2,4,5,9
2000: 2.5,7,7,7

So, last place varies from 1-4 wins. Average is about 3 wins.
Second-last varies from 1-7 wins. Average is about 4 wins.
Third-last varies from 4-7 wins. Average is a bit over 5 wins.
Fourth-last varies from 5-9 wins. Average is about 7 wins.

This year might be one of the most even on record, or it might just be that we don't yet know who the dunces are.

The squiggle will be conservative in its estimates for two reasons.
(1) The ratings are a bit tentative this early in the season. Once more results are in, the predictions will be more emphatic.
(2) As some teams' partial wins are converted to losses, the win total will drop. For these teams, their rating may also go down as this happens.