Certified Legendary Thread Race for the flag, in squiggly lines

[cryptor]

Sadly, I read all of what you said.

"North to me have been a bit like Freo last year to me."

Wrong. Don't care what you feel, your feelings are wrong as the teams have had nothing in common.

"Just average wins as you'd expect them to get."

That isn't how Fremantle played last year early in the season.

"Winning is great and you can only play who you've come up against so can't hold that against them. Just simply haven't seen anything from them that makes me think they have what it takes. Plenty of time for them to show it though."

We haven't played the good teams yet, so how those results go will largely determine what people feel, yet Fremantle won all of theirs, most quite convincingly, other than against Hawthorn. You are setting up a double negative scenario in which you have successfully mind-stuffed yourself to write the team off irrespective what occurs. If we don't beat any of the good teams, we are no good, if we drop any of them, then we are like Fremantle and are going to bomb out like they did and not win any games in 2017.

On the other hand...

"West Coast on the other hand have smashed some teams with big margins and their 3 losses while being bad were still against Hawthorn, Sydney and Geelong at each of those sides home ground"

Beaten nobody, struggle to win away from Subi against good opposition, smash shit sides at home... winning formula for success?

"Either way there's plenty more water to go under the bridge. North have to play more of the top 8 and Eagles have yet to play many top 8 sides on their deck."

That is the point, there is no water to go under the bridge, you have constructed a scenario where irrespective what we do we are in a no-win scenario because you have convinced yourself that we are the 2016 version of Fremantle, irrespective that there are absolutely no similarities. Fremantle basically only lost a handful and beat almost all of the contenders, what possible scenario would we not be Fremantle in your eyes?

The funny thing is I don't think you even consciously realise what you are doing.

I have no issue if people don't rate us, we aren't travelling that great tbh, I just don't think people need to fabricate reasons to write us off when there are plenty of legitimate reasons to do so.

Fremantle had some good performances early last year but their results weren't anything special.

R1 - 7 points over Port at home
R2 - 44 points over Geelong away (their bunnies and Geelong weren't top 8)
R3 - 30 points over Eagles
R4 - 14 points over Swans at home
R5 - 68 points over Melbourne away (who still hadn't come good)
R6 - 28 points over Bombers at home
R7 - 13 points over Dogs away
R8 - 73 points over North at home
R9 - 11 points over Adelaide away
R10 - Lost to Richmond at home by 27

Most all of those games were well within expectations for Fremantle. The only exception the game against North and perhaps that's the reason you've convinced yourself that Freo were far more dominant than they really were.

And I don't know where this BS about me setting up a lose lose situation for North comes from. There are another 14 regular season matches for North and all they need to do to get me to change my mind about them is to have some massive (10+ goals min) wins against the teams everyone expects them to beat and/or have some really convincing wins against other contenders at home (6+ goals) or simply win against those contenders away.

They haven't done any of that yet. But there's plenty of chance left for that to happen.

 

[Matele]

Have just become a Squiggle convert.
Squiggle is love. Squiggle is life.

One of us, one of us..

 

[Footy_Fan2007]

Why can't eagles have a go away from home?

They tried and got thumped by the cats.

 

[Final Siren]

I wonder if Final Siren could provide a "by team" home ground advantage factor, and also an away game factor. I'm sure the Eagles (for example) would have a huge difference in both directions.

Or even a by-team "accuracy" factor - which team is most consistent as per squiggle? Which is harder to pick?

West Coast have been the easiest to tip (8/8)... as you'd probably expect, since the Eagles have played weak teams at home and strong teams away.

Only 3/8 for Melbourne, Collingwood, Richmond and Port Adelaide, and only 4/8 for Geelong. Those last three are probably all attributable to the squiggle being fairly conservative and slower to throw away last year's form than most human punters.

Home Ground Advantage varies depending on time period and method, but I get something like the below. Coloured bars to the right of the 0 line mean a team overperforms there:the longer the bar, the more it overperforms. Bars to the left of the 0 line mean underperformance.

West Coast at Subi is indeed the biggest overperformer! By a fair margin, too. They don't vary much in other places, though.

Also notable:

• Hawthorn overperform almost everywhere
• Fremantle don't vary much
• The Bulldogs really like Docklands
• Richmond are terrible at their home ground

 

[Gavin Excell]

West Coast have been the easiest to tip (8/8)... as you'd probably expect, since the Eagles have played weak teams at home and strong teams away.

Only 3/8 for Melbourne, Collingwood, Richmond and Port Adelaide, and only 4/8 for Geelong. Those last three are probably all attributable to the squiggle being fairly conservative and slower to throw away last year's form than most human punters.

Home Ground Advantage varies depending on time period and method, but I get something like the below. Coloured bars to the right of the 0 line mean a team overperforms there:the longer the bar, the more it overperforms. Bars to the left of the 0 line mean underperformance.

West Coast at Subi is indeed the biggest overperformer! By a fair margin, too. They don't vary much in other places, though.

Also notable:

• Hawthorn overperform almost everywhere
• Fremantle don't vary much
• The Bulldogs really like Docklands
• Richmond are terrible at their home ground

Interesting re Tigers

 

[tazaa]

West Coast have been the easiest to tip (8/8)... as you'd probably expect, since the Eagles have played weak teams at home and strong teams away.

Only 3/8 for Melbourne, Collingwood, Richmond and Port Adelaide, and only 4/8 for Geelong. Those last three are probably all attributable to the squiggle being fairly conservative and slower to throw away last year's form than most human punters.

Home Ground Advantage varies depending on time period and method, but I get something like the below. Coloured bars to the right of the 0 line mean a team overperforms there:the longer the bar, the more it overperforms. Bars to the left of the 0 line mean underperformance.

West Coast at Subi is indeed the biggest overperformer! By a fair margin, too. They don't vary much in other places, though.

Also notable:

• Hawthorn overperform almost everywhere
• Fremantle don't vary much
• The Bulldogs really like Docklands
• Richmond are terrible at their home ground

Carlton sure do love the MCG. No wonder they want to get out of Etihad.
On another note, As long North get a home final at the G in September in the top four, il be happy that we can get a result.

 

[RedStarUncle]

Nice post, thanks for your time, but you're misunderstanding as well. I'm not challenging the tool, but the analysis/context/conclusions made from it. I'm rejecting the claim that holding Saints to three goals is anyway indicative of West Coast being a a GF team or that you can claim they are a better chance than North at this stage.

I would claim that Eagles v Saints results is indicative more that "Good teams keep bad teams to low scores" than the claim that it is proof of "Grand Final winners keep bad teams to low scores".

If you suggest the evidence points the other way then I'd be interested in seeing that data; over the past seasons which teams that hold their opponents to 4 or less goals in match, where they end up at seasons end.

If the result proves good teams keep bad teams to low scores, and if it's reasonable to presume that good teams win Grand Finals which seems quite intuitive, then by definition teams that keep bad teams to low scores are more likely to win Grand Finals than teams that don't.

History tells us that a team like the Dockers last year or Roos this year isn't likely to win a Grand Final. All those narrow wins against average sides eventually lead to losses in finals when you only play good sides. The real contenders, on the other hand, are the teams that are a class above the average sides and make them look bad. North didn't make St Kilda look bad but the Eagles did, so I think it of course makes sense that we should see West Coast as the more likely of the two to be a genuinely good side.



If your only doubt is that the Squiggle doesn't do a very good job of predicting the likelihood of an interstate team with a poor record at the MCG beating a Melbourne-based team in a Grand Final, then maybe that's a fair doubt but I'd have to let the real experts answer the point.

 

[RedStarUncle]

That's because it was nonsense



I'm not attacking it, merely questioning where the teams are being rated and how they're faring in relation to the ratings. I don't think Squiggle predicted Carlton to win 4 in a row, or Collingwood to be so bad, or North to be 8-0. Squiggles had Radelaide pegged as almost grand final certainties at one point and West Coast as dead set grand finalists. It certainly didn't rate Geelong or Sydney and was no where near Melbourne, GWS or Fremantle (but then again, I don't think anyone was).

I like the Squiggles, but I have a questioning nature and I don't take everything at face value. Besides, feedback provides better tools and questioning the methods to obtain a result generally leads to improved methods.

I think that this season is a far different season to any in recent memory and teams are performing unexpectedly worse, or better because of it, and it could very well be that the ratings may need to adjust to work with the changes. They're substantial interpretation and rule changes, not some simple holding the ball adjustment.

This is why the Squiggle is great... once those results happened, we got an updated Squiggle with an updated impression of how the season will unfold and a new perspective on how 2016 is tracking.

Unlike media experts and special comments types (or even your average fan), Squiggle doesn't hold onto the pre-conceived notions it had at the start of the year... it updates itself constantly with each result. If Carlton turn their 4-game winning streak into an all-time shock Premiership, Squiggle will be there telling the story of how they rose and rose and rose to that lofty standard defying early predictions.

 

[SJ]

West Coast have been the easiest to tip (8/8)... as you'd probably expect, since the Eagles have played weak teams at home and strong teams away.

Only 3/8 for Melbourne, Collingwood, Richmond and Port Adelaide, and only 4/8 for Geelong. Those last three are probably all attributable to the squiggle being fairly conservative and slower to throw away last year's form than most human punters.

Home Ground Advantage varies depending on time period and method, but I get something like the below. Coloured bars to the right of the 0 line mean a team overperforms there:the longer the bar, the more it overperforms. Bars to the left of the 0 line mean underperformance.

West Coast at Subi is indeed the biggest overperformer! By a fair margin, too. They don't vary much in other places, though.

Also notable:

• Hawthorn overperform almost everywhere
• Fremantle don't vary much
• The Bulldogs really like Docklands
• Richmond are terrible at their home ground

This is excellent. Can you break this into home/home state/away state?

 

[Crankyhawk]

Oh dear.

First the Power Rankings, now The Squiggle, Chunkchicken getting owned by all the rankings systems.

Oi! Chunkychicken smashed Roby from pillar to post...

 

[Simon_Nesbit]

Just trying to wrap my head around it - are these bars cumulative measures or distinct?

To use Hawthorn as an example - does it mean Hawthorn outperform expectations by ~55% at MCG, and ~45% at Docklands, or rather are they only figures of say ~15% and ~5%?

also where is York Park (Aurora Stadium in Launceston)?

 

[Bloody]

Final siren the out performs? Is that points? Or percentage above opponents?


Sent from my iPhone using Tapatalk

 

[HawkintheTuarts]

This is just evidence from 2 seasons but many seasons are like this which is why the flagpole has an 80% hit rate at predicting winners in the last 20 years.

Nah, it's cherry-picking, not evidence, to suit an argument. I could do the same. Eg last year Melbourne holding Brisbane to 4 goals, or in 2014 GWS holding Melbourne to 4 goals. When Adelaide held Fremantle to just one goal in '09, they only made it to the Semi's, not the GF. Those are examples, not all the evidence.

 

[BEaston]

^ no offence but you obviously don't understand how the flagpole works. It's a model based on historical data, tested via simulation. I suggest stop posting and read up about it on the site.

 

[djpaulydemon]

Nah, it's cherry-picking, not evidence, to suit an argument. I could do the same. Eg last year Melbourne holding Brisbane to 4 goals, or in 2014 GWS holding Melbourne to 4 goals. When Adelaide held Fremantle to just one goal in '09, they only made it to the Semi's, not the GF. Those are examples, not all the evidence.

You lack basic comprehension. I didn't say it was all the evidence and i'm not trying to convince you that holding a team to a few goals once will win you a premiership.

Thrashing multiple teams is the indicator. Hollding teams to a low score is a by-product. Teams that do it the most (winning by a lot) in a season usually win the flag.

 

[yoodooright]

You lack basic comprehension. I didn't say it was all the evidence and i'm not trying to convince you that holding a team to a few goals once will win you a premiership.

Thrashing multiple teams is the indicator. Holding teams to a low score is a by-product. Teams that do it the most (winning by a lot) in a season usually win the flag.

I think his point may be that correlation =/= causation necessarily. Nothing wrong with questioning the model. But yes, there does seem to be a strong correlation between utterly dominant teams & flags won.

I dunno though, it is hard to follow his argument.

 

[Lsta062]

Interesting re Tigers

It is a problem that we noticed for a couple of seasons now.

We play better against some teams interstate than we do at the MCG, like Brisbane, GWS, Fremantle and Port Adelaide (Funny that those sides have the MCG as one of the stadiums they overperform in). I don't see us having much of an advantage over sides at the MCG apart from the teams that struggle at the MCG like Sydney, Adelaide, Gold Coast etc.

If you can play all around Australia, but your weakest ground is your home ground, then you are limited in how good you can be. It is a problem that needs to be rectified.

 

[philohk]

It is a problem that we noticed for a couple of seasons now.

We play better against some teams interstate than we do at the MCG, like Brisbane, GWS, Fremantle and Port Adelaide (Funny that those sides have the MCG as one of the stadiums they overperform in). I don't see us having much of an advantage over sides at the MCG apart from the teams that struggle at the MCG like Sydney, Adelaide, Gold Coast etc.

If you can play all around Australia, but your weakest ground is your home ground, then you are limited in how good you can be. It is a problem that needs to be rectified.

Tiges are the bizarro Eagles.

 

[Final Siren]

Just trying to wrap my head around it - are these bars cumulative measures or distinct?

To use Hawthorn as an example - does it mean Hawthorn outperform expectations by ~55% at MCG, and ~45% at Docklands, or rather are they only figures of say ~15% and ~5%?

also where is York Park (Aurora Stadium in Launceston)?

Only the size of each distinct color block matters, not where it's positioned (aside from whether it's left or right of the 0 line).

The units are points, i.e. West Coast are currently a 47-point better team at Subiaco, according to this metric. This works by measuring the difference between actual and expected scores and adding some of the difference to that team's "home ground advantage" counter. ("Some" here is 87% of the square root.)

The sample size can be very small in some cases, since some teams play only rarely on some grounds. And even with home grounds, the advantage fluctuates a lot over the years, as teams deliver unexpected results. In fact, some "home ground advantage" measured here is probably just teams getting better or worse before the algorithm has had a chance to figure out that this is their new normal, and nothing to do with the ground at all.

But anyway! It does more or less line up with what you'd expect.

West Coast had a run of bad results at Subi in 2013 but have gotten ever-stronger since:

 

[beert]

West Coast have been the easiest to tip (8/8)... as you'd probably expect, since the Eagles have played weak teams at home and strong teams away.

Only 3/8 for Melbourne, Collingwood, Richmond and Port Adelaide, and only 4/8 for Geelong. Those last three are probably all attributable to the squiggle being fairly conservative and slower to throw away last year's form than most human punters.

How does it go on margins though? Which teams have the highest standard deviation for difference between squiggle tips and actual scores?

 

[The Swans Blog]

Only the size of each distinct color block matters, not where it's positioned (aside from whether it's left or right of the 0 line).

The units are points, i.e. West Coast are currently a 47-point better team at Subiaco, according to this metric. This works by measuring the difference between actual and expected scores and adding some of the difference to that team's "home ground advantage" counter. ("Some" here is 87% of the square root.)

The sample size can be very small in some cases, since some teams play only rarely on some grounds. And even with home grounds, the advantage fluctuates a lot over the years, as teams deliver unexpected results. In fact, some "home ground advantage" measured here is probably just teams getting better or worse before the algorithm has had a chance to figure out that this is their new normal, and nothing to do with the ground at all.

But anyway! It does more or less line up with what you'd expect.

West Coast had a run of bad results at Subi in 2013 but have gotten ever-stronger since:

I'd expect to see a very similar drop for points away from home.

 

[Simon_Nesbit]

Only the size of each distinct color block matters, not where it's positioned (aside from whether it's left or right of the 0 line).

Thanks, that makes perfect sense.

...but where is Aurora?

 

[Final Siren]

Thanks, that makes perfect sense.

...but where is Aurora?

Oh it fell off the graph somehow, sorry. Hawks are +11.5, GWS, Sydney, and St Kilda are all about +6, North are -8 and Freo are -15. Everyone else is 0 or thereabouts.

 

[Final Siren]

How does it go on margins though? Which teams have the highest standard deviation for difference between squiggle tips and actual scores?

You can tell that by eyeballing the chart. Teams with a lot of movement (e.g. Geelong, GWS) have had games with unexpected results, while not much movement (e.g. West Coast, Essendon, Freo after R2) means scores were close to prediction.

 

[Blue and Silver]

Quick question. Has the squiggle taken into account the Sam Kerridge vs North effect?

An extra 6 goals is nothing to sneeze at