Certified Legendary Thread Race for the flag, in squiggly lines

[sproketman]

The statistics of points for/against are used in the squiggle. The flagpole algorithm i'm sure uses a somewhat similar mechanism based on game stats to generate its data.

You can be semantic about the terms I used if you want. In any case my point remains - just because YOU don't agree with the specific algorithm used because it denigrates your team in some way doesn't make it 'useless'.
I personally find it quite useful, and would do so regardless of where Hawthorn is placed.

The statistics AND assumptions made to weight those statistics. Each model has a different set of assumptions to best fit the data. It's those assumptions I'm challenging.

Let's look at it another way. Hawthorn are ranked well above West Coast on flagpole, but they have to play West Coast away from home in the first final. It's close to a 50/50 game but the Squiggle predicts a loss for Hawthorn.

That means Hawthorn would have to play two games before a GF, while West Coast have to play one.

The permutations involved for Hawthorn to win the premiership would then be more difficult than West Coast's, as they have to win three games consecutively and travel once more, versus West Coast's two games, a rest, and only travelling for the GF.

Sydney by the same token are ranked well below both Hawthorn and West Coast, but if they win their QF get a week off and face one of them at home. And based on their current trajectory of form, would be expected to win against either opponent, or at least give a better contest than Freo.

Where did the flagpole rate Sydney against Hawthorn after the last round of 2012? Or Geelong against Collingwood in the second to last round of 2011? Or Collingwood against Geelong after the last round of 2010?

 

[sproketman]

Or Collingwood against Geelong after the last round of 2010?

I have my answer.

So as you can see, flagpole predicting premiers before finals is sub optimal, because a single result (St Kilda upsetting Geelong in the QF) left it difficult for the Cats to win.

That's not even one of the better cases of flagpole missing the premiers in the final rounds.

 

[iluvparis]

It's not supposed to have a 100% prediction record.

 

[Janus]

The flagpole also says the Crows are a better chance of winning a premiership than Fremantle, so I don't know why this guy is worried about us for...

Anyway, it seems like we've just gone around in circles this year. Almost right back where we started. Figures.

 

[Final Siren]

[...] Individual games are picked by standard Squiggle algorithm, the grand final is pick by its own separate algorithm than rates recent form higher, and flagpole is picked by a third unrelated one that rates scoring more favourably. [...]

Not quite; the flagpole algorithm and the GF tipper are the same thing.

 

[Royce Hafey]

I like how Hawthorn are in 'Hawthorn territory' West Coast are in 'West Coast Territory' and now Sydney is in 'Sydney Territory'.

Freo have not won a flag yet, maybe they are locking up their own little piece of premiership squiggle territory.

Richmond in their own new world as well. Even if you go back to the Hafey era - those far more were attacking teams then Dimma's mob.

 

[The Swans Blog]

I have my answer.

So as you can see, flagpole predicting premiers before finals is sub optimal, because a single result (St Kilda upsetting Geelong in the QF) left it difficult for the Cats to win.

That's not even one of the better cases of flagpole missing the premiers in the final rounds.

And that the pies absolutely thrashed them in the PF.

 

[armpit]

You can see what's going on clearly this week with freo. They have cleared space and are going to rocket up the open corridor.

 

[sproketman]

It's not supposed to have a 100% prediction record.

I never said it had to be, and that is not the issue. But it should at least be internally consistent according to its own logic.

For instance, the flagpole currently rates Adelaide as closer to the flag than Sydney, but Sydney has a much easier path to the flag. Adelaide is predicted to lose its first final by 8 points, Sydney is predicted to lose theirs by 1 point. The model only has to be slightly wrong for Sydney to go through to the prelims, while it has to be very wrong (two consecutive upsets) for Adelaide to do the same. Every other result can stay the same, and Sydney just about become favourites in both their remaining games after the qualifying final.

(Sydney are rated as a 1 point worse team than Freo, Hawthorn an 11 point better team than the Dockers, meaning the difference between the two is two goals, the difference that comes with home state advantage. If Sydney host the Hawks in a prelim it is likely they will be slight favourites against Hawthorn, and possibly slight favourites against West Coast in the GF, given their 'recent form' will arguably be stronger)

So the model doesn't have to be drastically wrong in its game-by-game tipping to be drastically wrong overall. Sydney can conceivably win the premiership and still be fourth in the flagpole.

I don't disagree with the flagpole's forecast that Hawthorn and West Coast are clear one and two best teams, but the gap between them and the next best is not as big as that graphic makes out.

And that the pies absolutely thrashed them in the PF.

But that came well after the flagpole's prediction in the last round. Three games after in fact.

 

[poshman]

The most interesting part of this to me is that not only do the Hawks have to leave Vic to play week one. They may have to do it twice, the second of which would (most likely) be against a rested side in Perth.

I'm not sure what the stars are but how often does a team win the prelim away from home and interstate against a rested side?

The Hawks are a great team but that would require great luck with fitness/health and an amazing result just to get to the GF. Not sure the squiggle in all its goodness can/has factored that in.

 

[Elite]

Nice.

I thought I'd have a look at 1990

Check the Pies out from round 18 onwards......

http://maxbarry.com/squiggle/1990.html

and in 2011, shouldve been back to back.... clearly an awesome squiggle lol.... unfortunately everything to could go wrong, went wrong.

 

[The Swans Blog]

The most interesting part of this to me is that not only do the Hawks have to leave Vic to play week one. They may have to do it twice, the second of which would (most likely) be against a rested side in Perth.

I'm not sure what the stars are but how often does a team win the prelim away from home and interstate against a rested side?

The Hawks are a great team but that would require great luck with fitness/health and an amazing result just to get to the GF. Not sure the squiggle in all its goodness can/has factored that in.

Sydney 2005.

 

[holybishop]

The most interesting part of this to me is that not only do the Hawks have to leave Vic to play week one. They may have to do it twice, the second of which would (most likely) be against a rested side in Perth.

I'm not sure what the stars are but how often does a team win the prelim away from home and interstate against a rested side?

The Hawks are a great team but that would require great luck with fitness/health and an amazing result just to get to the GF. Not sure the squiggle in all its goodness can/has factored that in.

Sydney 2005.

West Coast 2006 as well. That's the last time a team has won a prelim after losing week 1 though.

 

[The Swans Blog]

West Coast 2006 as well. That's the last time a team has won a prelim after losing week 1 though.

They played at home in 2006 then away in 2005. Swans played 3 away games that finals series and still came up with the goods.

 

[Final Siren]

Where did the flagpole rate Sydney against Hawthorn after the last round of 2012?

Or Geelong against Collingwood in the second to last round of 2011?

Or Collingwood against Geelong after the last round of 2010?

I have my answer.

So as you can see, flagpole predicting premiers before finals is sub optimal, because a single result (St Kilda upsetting Geelong in the QF) left it difficult for the Cats to win.

That's not even one of the better cases of flagpole missing the premiers in the final rounds.

I'd actually say St. Kilda is the one to criticize there. Collingwood being rated a high 2nd a month before they finished 1st is pretty good. But St. Kilda are low and came very close to winning a Grand Final.

 

[Hawkk]

I'd actually say St. Kilda is the one to criticize there. Collingwood being rated a high 2nd a month before they finished 1st is pretty good. But St. Kilda are low and came very close to winning a Grand Final.

The 2012 squiggle hurts. That said we lost Brent Guerra in the last round of the season and our defensive structures were all array come finals time

 

[RedmanWasHere]

If I'm looking at it right, is the height of the flagpole proportionate to how attacking and adept at defending a side is?

 

[iluvparis]

If I'm looking at it right, is the height of the flagpole proportionate to how attacking and adept at defending a side is?

Yes - it ignores things like where finals would be played and what position teams currently are in the ladder (not sure everyone gets those latter points).

 

[poshman]

Sydney 2005.

So they need to do something that's been done once.

I also tyinl the competition is more even this year than 05.

Big ask!

 

[Golumless]

Nice.

I thought I'd have a look at 1990

Check the Pies out from round 18 onwards......

http://maxbarry.com/squiggle/1990.html

and in 2011, shouldve been back to back.... clearly an awesome squiggle lol.... unfortunately everything to could go wrong, went wrong.

That is a nice squiggle, but, Collingwood 1990 has nothing on Fitzroy 1899.

http://maxbarry.com/squiggle/1899.html

Edit: 1899 proved to be much more interesting than 1898.

 

[FootyTradePro]

The thing I love about the old squiggle is you can tell when there was a really wet and rainy Saturday.

All the teams move to the right in unison!

 

[Dr Tigris]

Is my, and many others, feeling that this is an unusually even year true by the squiggle? Just looking at the squiggle/flagpole it doesn't seem so even. But is sure feels that way week to week.

 

[old55]

These are not statistics and probabilities. It's a number of heuristic models that change depending on their best fit for the result. Individual games are picked by standard Squiggle algorithm, the grand final is pick by its own separate algorithm than rates recent form higher, and flagpole is picked by a third unrelated one that rates scoring more favourably.

The only algorithm that would make sense in terms of 'who is closest to the flag' would be one based on a Markov chain. Based on the Squiggle predictions for games, Hawthorn and West Coast are reasonable chances of not making the GF, if they didn't, would they still lead in the flag pole?

I'm no maths wizz (but I did know the bucket cost 5c) so I'm happy to be corrected, but I don't think you're looking for a Markov chain because I think that relies on a standard transition matrix, this scenario doesn't have that - the transition probabilities vary from round to round.

Like I've said before - the biggest value in the Squiggle for me is data visualisation.

 

[The Swans Blog]

I never said it had to be, and that is not the issue. But it should at least be internally consistent according to its own logic.

For instance, the flagpole currently rates Adelaide as closer to the flag than Sydney, but Sydney has a much easier path to the flag. Adelaide is predicted to lose its first final by 8 points, Sydney is predicted to lose theirs by 1 point. The model only has to be slightly wrong for Sydney to go through to the prelims, while it has to be very wrong (two consecutive upsets) for Adelaide to do the same. Every other result can stay the same, and Sydney just about become favourites in both their remaining games after the qualifying final.

(Sydney are rated as a 1 point worse team than Freo, Hawthorn an 11 point better team than the Dockers, meaning the difference between the two is two goals, the difference that comes with home state advantage. If Sydney host the Hawks in a prelim it is likely they will be slight favourites against Hawthorn, and possibly slight favourites against West Coast in the GF, given their 'recent form' will arguably be stronger)

So the model doesn't have to be drastically wrong in its game-by-game tipping to be drastically wrong overall. Sydney can conceivably win the premiership and still be fourth in the flagpole.

I don't disagree with the flagpole's forecast that Hawthorn and West Coast are clear one and two best teams, but the gap between them and the next best is not as big as that graphic makes out.

But that came well after the flagpole's prediction in the last round. Three games after in fact.

If Sydney & Hawthorn win, then it's likely Eagles vs Crows out west, which Eagles should win, and Freo hosting Richmond, which they should win as well. Otherwise if Sydney and West Coast win, then there's a PF showdown and Sydney wouldn't be favourites against Hawthorn.

The most ideal path for Sydney and Hawthorn to the grand final is one where they don't cross each others path!

Round 1:
Hawthorn def. Eagles
Sydney def. Fremantle
Bulldogs def. Crows
North def. Richmond

Round 2:
Eagles def. Dogs
Richmond def. Freo

Round 3:
Swans def. Eagles
Hawks def. Richmond

 

[ep2018]

They played at home in 2006 then away in 2005. Swans played 3 away games that finals series and still came up with the goods.

That's why I love the current finals format so much. That we could fluff up our first final as minor premier and still be a chance to win the flag.