Certified Legendary Thread Squiggle 2017

[Bloody]

Careful, try not to let your bias show.

The username and post made me chuckle


Sent from my iPhone using Tapatalk

 

[sthmelb_dimmies]

Is there a way to have videos permanently removed from the Internet

I'm asking for a friend...

Sure. Report it for copyright violation.

 

[Juke]

Richmond's flagpole does not do anything to dissuade me from thinking that we're on track to losing our 4th EF in a row. Our scoring power is still very low, despite looking better at the start of the season.

 

[chickentendies]

Good times.



And 2005 was 7 years after your last grand final appearance.

Can we not turn this thread into a West Horsham shitfight plz

 

[Dirty Bird]

That was effectively 1st v 8th though. Think teams ate far more evenly matched now.

Stop playing the underdog role
I could do the same

Wooden spoon favourites vs reigning premiers

 

[Final Siren]

Gotta love how gws and hawthorns flag pole predictor goes up after their loses. Becomes almost impossible now for hawthorn to make finals and it goes up. GWS lose to a lower ranked team and it goes up. It is so stupid.

GWS lost to a higher ranked team (Sydney). But I agree it's misleading to have a "Premier Predictor" that rates teams who can't make finals higher than zero.

Now where it gets interesting is when you have to decide what to do about teams who might make finals. My first thought was to downgrade them based on their chance of making it, but I discovered this starts to work against the whole point of what Flagpole is supposed to do, which is to detect flag-winning form even when the team in question isn't sitting in the Top 2. Instead it becomes a rating of who's most likely to finish in the top ladder spots, which we already knew. It also depresses the rating of teams like the Bulldogs 2016 and Sydney 2014 (and Sydney 2017?), who at various stages had a real risk of missing finals but then made it and went on to have a successful campaign, while benefiting a team like Fremantle 2015, who were never going to do well in finals, but were at least certain of making it in the door.

And it creates very large movements when a team seems likely to miss finals but unexpectedly turns things around, like Richmond 2014. Or vice versa, like Port Adelaide 2015. This would be okay in a year like 1993 (ahhhh, 1993) where Geelong were a serious flag chance but then didn't quite make it, but under normal circumstances, I don't know if I want borderline teams yoyoing all over the place.

It does do all right in years like 2015, when Hawthorn and West Coast were the clear top 2 both in terms of flag-winning form and likely ladder position. But then Flagpole already gets those years right, so there's no real gain.

Anyway there will be a future version of Flagpole that pushes teams down based on their chances of making finals, but it's reasonably tricky to get right and not ready yet.

 

[Hobbes]

Yes the maths is stupid. Its hocus pocus statistics the lures in the naive. If offends me like natropaths offend doctors.

The predictive value has been demonstrated. It's not just woo.

 

[WhiskeyTangoFoxtrot]

Yes the maths is stupid. Its hocus pocus statistics the lures in the naive. If offends me like natropaths offend doctors.

By all means, please show us your statistical model.

 

[fishardansin]

Final Siren if I may ask a few Essendon focused questions:

Does the fact that we only have one more game against a team above mean that there is not much chance of significant improvement in Squiggle left? For example, if Essendon won all games except Adelaide would that see any growth at all or just be expected at this point?

 

[WhiskeyTangoFoxtrot]

Final Siren if I may ask a few Essendon focused questions:

Does the fact that we only have one more game against a team above mean that there is not much chance of significant improvement in Squiggle left? For example, if Essendon won all games except Adelaide would that see any growth at all or just be expected at this point?

I'm not Final Siren, but it will come down to how much you win those matches. If you win better than the Squiggle expects you to then you will Squiggle upwards.

If you lose, or win by less than the Squiggle expects then you will Squiggle downwards.

 

[less than zero]

Why has it taken me so long to discover this thread! I love it........mesmerizing.

 

[Headless]

Why has it taken me so long to discover this thread! I love it........mesmerizing.

You'll simply never leave. Welcome.

 

[Forward Press]

Why has it taken me so long to discover this thread! I love it........mesmerizing.

Thread is like the Hotel California - you can check out (other threads) any time you like, but you can never leave.

 

[less than zero]

You'll simply never leave. Welcome.

I'm looking forward to the round being over before it's started......the power of squiggle!

 

[Dave]

How have I only just discovered this, ripper of a thread

Need to venture out of the bay more m8

 

[craffles]

Need to venture out of the bay more m8

I reckon, if I post a terrible photoshop on the bombers board I rack up bulk uncontested likes m7

 

[less than zero]

Thread is like the Hotel California - you can check out (other threads) any time you like, but you can never leave.

Thread needs Hotel California like promoting, only 42 pages since Feb, means bigfooty posters are being denied, poor souls. An Eagles fan might be able to assist with this.

 

[Forward Press]

Thread needs Hotel California like promoting, only 42 pages since Feb, means bigfooty posters are being denied, poor souls. An Eagles fan might be able to assist with this.

Little known fact - 'Life in the Fast Lane' was a satirical piece written by Glenn Frey about Matt Priddis.

 

[GreyCrow]

Thread is like the Hotel California - you can check out (other threads) any time you like, but you can never leave.

Her mind is Squiggly-twisted, she got the Mercedes bends
She got a lot of pretty, pretty boys she calls friends
How they post in the courtyard, sweet summer sweat.
Some read to remember, some post to forget

 

[sthmelb_dimmies]

Final Siren can you walk me through the var = towerdata [] dataset? This is what I assume is for Round 18

If I was to say, I dunno, work out what % chance Richmond have of finishing say, 9th, how do I use the integers in their row?

 

[sthmelb_dimmies]

Nah, got it. So each number is the x-axis width of the team color at position n. For total row width of 100,000.

So Tower currently reckons Richmond finishing 9th is 14.743%

 

[Randomizor]

GWS lost to a higher ranked team (Sydney). But I agree it's misleading to have a "Premier Predictor" that rates teams who can't make finals higher than zero.

Now where it gets interesting is when you have to decide what to do about teams who might make finals. My first thought was to downgrade them based on their chance of making it, but I discovered this starts to work against the whole point of what Flagpole is supposed to do, which is to detect flag-winning form even when the team in question isn't sitting in the Top 2. Instead it becomes a rating of who's most likely to finish in the top ladder spots, which we already knew. It also depresses the rating of teams like the Bulldogs 2016 and Sydney 2014 (and Sydney 2017?), who at various stages had a real risk of missing finals but then made it and went on to have a successful campaign, while benefiting a team like Fremantle 2015, who were never going to do well in finals, but were at least certain of making it in the door.

And it creates very large movements when a team seems likely to miss finals but unexpectedly turns things around, like Richmond 2014. Or vice versa, like Port Adelaide 2015. This would be okay in a year like 1993 (ahhhh, 1993) where Geelong were a serious flag chance but then didn't quite make it, but under normal circumstances, I don't know if I want borderline teams yoyoing all over the place.

It does do all right in years like 2015, when Hawthorn and West Coast were the clear top 2 both in terms of flag-winning form and likely ladder position. But then Flagpole already gets those years right, so there's no real gain.

Anyway there will be a future version of Flagpole that pushes teams down based on their chances of making finals, but it's reasonably tricky to get right and not ready yet.

Making it to prelim = not doing well in finals?

You must hate supporting Richmond then

 

[Final Siren]

Making it to prelim = not doing well in finals?

You must hate supporting Richmond then

Fremantle made a prelim in 2015 because:
(a) They were a genuinely good team for the first 9 rounds
(b) They fell over the line in an incredible number of close games thereafter
(c) Teams competing for the Top 4 won & lost in an almost eerily perfect pattern that enabled Fremantle to finish on top
(d) They had a final at home against a badly injured Sydney who kicked 7.18

 

[Final Siren]

Nah, got it. So each number is the x-axis width of the team color at position n. For total row width of 100,000.

So Tower currently reckons Richmond finishing 9th is 14.743%

Almost - the first and last array elements are repeated to force the graph to display something at the very top and bottom. Richmond are 13.1% chance of 9th.

Also something to be aware of is that when it auto-generates sims mid-round, there are only 2,500 total, not 100,000.

 

[Final Siren]

Why has it taken me so long to discover this thread! I love it........mesmerizing.

After doing this for a few years, I've noticed a suspicious correlation between the number of each team's supporters in this thread and how well that team is rated on Squiggle.

I miss the Hawthorn guys. They were quietly smug and genuinely curious to analyze how great their team was.

Fremantle fans used to be everywhere, before the squiggle turned.

West Coast fans didn't believe what was happening almost until they made the finals but then went away again.

Adelaide fans are numerous nowadays but wary after past unfulfilled promises.