- Apr 4, 2013
- 13,670
- 16,890
- AFL Club
- Fremantle
Anyway, Fox Sports and Champion Data have stolen Final Siren's IP, no one is stealing Roby's. I think that explains enough. Power Rankings: Too weird to live and too rare to die.
Certified Legendary Thread Race for the flag, in squiggly lines
- Thread starter Final Siren
- Start date
- Tagged users None
I made money on Fremantle today.
- Apr 4, 2013
- 13,670
- 16,890
- AFL Club
- Fremantle
Yeah but how many beans.
Spoken4
Norm Smith Medallist
Roby, you predicting that Freo won't make it to the grand final makes me surer that Freo will.
Your predictions simply have no credibility whatsoever whereas the squiggle is actually based on logical and substantiated assumptions.
I will continue to make money on Fremantle for the rest of the year while you have to watch the Eagles play in a Grand Final.
Good luck.
- Apr 4, 2013
- 13,670
- 16,890
- AFL Club
- Fremantle
After your complete meltdown in your own thread, I doubt it. The term 'cautionary tale' now has a synonym that is also a neologism: Roby.
The King!
Chosen One
Someone trying to derail this thread then
You reckon Final Siren's the dick?
- Oct 29, 2012
- 4,565
- 7,667
- AFL Club
- Hawthorn
- Other Teams
- Parramatta Eels, Arsenal, The Exers
I thought Roby had fecked off to write a book?
Why is he hanging around this thread like a bad smell?
Why is he hanging around this thread like a bad smell?
- Banned
- #3,486
Exactly, one bloke is a genuine stats nerd, happy to follow logic and when he does call others out does it with class and style.
The other bloke is a person with a serious gambling problem who needs genuine help and a 'model' based on subjective data and when he does call out others does it with a lack of class and style.
I have to respond.
Can you point to final siren being a dick to you? I'm genuinely curious whether this is being a dick or just constructive criticism.
This thread does a number of things:
1. It tells you who is most closest to the flag
2. It tells you the balance of all the sides, whether they're more attacking or defensive
3. It tells you how well the defense and offense of the teams are going and tells you in a much more objective manner than the eyeball test
4. It provides you with a generally accurate depiction of who are the better sides and who are the worse sides
None of these are tipping or betting, it's not even the main aim of the thread, which is called Race for the flag, in squiggly lines it says in the title that this is a thread designed to predict who will win the flag. Ultimately, there's only one tip that this thread cares about, and it's the last and most important one.
Final siren has posted his rate on predicting grand finalists and preliminary and semi finalists. It's pretty impressive. For the record Sydney were favourites for a lot of models, champion data predicted a Sydney win, the squiggle predicted a Sydney win by 10, footy maths predicted Hawthorn by one point and footy forecaster predicted a Hawthorn win by 11 points. Nearly every model was at least 40 points off the margin.
There's that, there's also St Kilda's wooden spoon season in 2014 where they conceded the most points of any side scored the second least amount of points, and the fact that Essendon only lost to Sydney in Sydney by 12 points.
So form matters right? Well footy maths factors in form and it's tipped 70 and the squiggle has tipped 71. Factoring in injuries opens up a pandoras box of problems that you've experienced yourself, how do you determine the WAP (wins above replacement) rating of a player given this is a game where individual performances can't be separated? You end up not knowing the true value of the replacements and plus, you never know the true value of the stars because this is a league which is never devoid of injuries. You want to say that Essendon play better with Carlisle? Go ahead, you're right. You want to create a power rankings system which accurately bakes injuries into its system and works? I've never seen one which actually functionally works, and that counts yours. You haven't statistically proven that Gold Coast is the best side in a league without injuries because there is no data for that in the first place.
As I said, both systems don't peg Gold Coast right, which is not surprising, they're a young side, no-one really knows their true weight.
So why are you abusing final siren for not tipping well then?
Humans are wonderfully clever, inventive and can think of ideas and create things that robots currently can't. However, humans are bad at recognizing patterns at a macro level without actually number crunching, you know, relying on machines to get some statistical analysis. I'm surprised a man of calculus is heckling a man who relies on data and cold hard numbers. Unless your system is, as final siren thinks, purely you sitting on the couch, there has to be some statistical basis for your system. In fact, given your lists about injury and umpiring influences and their statistical weightings, surely there is.
I like this thread. I liked yours. Both have limitations, final siren knows the limitations of the squiggle, do you know the limitations of your model?
Can you point to final siren being a dick to you? I'm genuinely curious whether this is being a dick or just constructive criticism.
This thread does a number of things:
1. It tells you who is most closest to the flag
2. It tells you the balance of all the sides, whether they're more attacking or defensive
3. It tells you how well the defense and offense of the teams are going and tells you in a much more objective manner than the eyeball test
4. It provides you with a generally accurate depiction of who are the better sides and who are the worse sides
None of these are tipping or betting, it's not even the main aim of the thread, which is called Race for the flag, in squiggly lines it says in the title that this is a thread designed to predict who will win the flag. Ultimately, there's only one tip that this thread cares about, and it's the last and most important one.
Final siren has posted his rate on predicting grand finalists and preliminary and semi finalists. It's pretty impressive. For the record Sydney were favourites for a lot of models, champion data predicted a Sydney win, the squiggle predicted a Sydney win by 10, footy maths predicted Hawthorn by one point and footy forecaster predicted a Hawthorn win by 11 points. Nearly every model was at least 40 points off the margin.
There's that, there's also St Kilda's wooden spoon season in 2014 where they conceded the most points of any side scored the second least amount of points, and the fact that Essendon only lost to Sydney in Sydney by 12 points.
So form matters right? Well footy maths factors in form and it's tipped 70 and the squiggle has tipped 71. Factoring in injuries opens up a pandoras box of problems that you've experienced yourself, how do you determine the WAP (wins above replacement) rating of a player given this is a game where individual performances can't be separated? You end up not knowing the true value of the replacements and plus, you never know the true value of the stars because this is a league which is never devoid of injuries. You want to say that Essendon play better with Carlisle? Go ahead, you're right. You want to create a power rankings system which accurately bakes injuries into its system and works? I've never seen one which actually functionally works, and that counts yours. You haven't statistically proven that Gold Coast is the best side in a league without injuries because there is no data for that in the first place.
As I said, both systems don't peg Gold Coast right, which is not surprising, they're a young side, no-one really knows their true weight.
So why are you abusing final siren for not tipping well then?
Humans are wonderfully clever, inventive and can think of ideas and create things that robots currently can't. However, humans are bad at recognizing patterns at a macro level without actually number crunching, you know, relying on machines to get some statistical analysis. I'm surprised a man of calculus is heckling a man who relies on data and cold hard numbers. Unless your system is, as final siren thinks, purely you sitting on the couch, there has to be some statistical basis for your system. In fact, given your lists about injury and umpiring influences and their statistical weightings, surely there is.
I like this thread. I liked yours. Both have limitations, final siren knows the limitations of the squiggle, do you know the limitations of your model?
Final Siren below.
The Rankings did predict at one stage they were more than 50% chance of making despite never been odds on. Just twisting facts.
http://www.bigfooty.com/forum/threads/2015-afl-power-rankings.1091941/page-34#post-38628605
And that's 2015.
2014 was no better.
Final Siren posted the following after I was seriously sick in hospital.
Yes very funny and empathetic.
- Apr 4, 2013
- 13,670
- 16,890
- AFL Club
- Fremantle
Poor lil fella.
See above.
1. It's a representation of an idea of who is closest to the flag. It's not actual reality. At best it does just as good a job as the bookmakers, never better.
2. Well yes it does it that.
3. What is this eyeball test?
4. Are the Bulldogs, Melbourne and St Kilda as bad as the Squiggle suggests?
And yet most of the talk is about tipping and forecasting. If it's not about those these why do those things get talked about the most.
Most models predicted Hawthorn or had Hawthorn on top.
Yes and it was a similar issue I had with my model which I needed to improve.
Other rankings systems around the world do so. I checked some NBA rankings recently and they were accounting for it.
Where is the abuse. I just said if you tip against it will make you money. How is that abuse? Did you read my previous post. I actually like checking the Squiggle. I don't like how Final Siren acts, belittling.
Yes there are many limitations to my model, many more than this one, due to the fact it uses more parameters. You will find in my posts I have mentioned limitations over and over, at ad nauseam. I do know why you haven't seen me post them because they are definitely there.
I'm only trying help you mate, I'm sorry about your Freo. Pavlich used to be my favourite player for a very long time and I always wanted him to win one.
Don't you have your own thread to talk about these kind of things?
FFS, everyone loves this thread and the squiggle. Why is this guy so salty? If you don't like it, and you've said your piece then stop being such an attention whore.
- Apr 4, 2013
- 13,670
- 16,890
- AFL Club
- Fremantle
Roby flicked the bean.
His thread was locked because his ultimate troll was eventually up and his last rants turned personal.
He has been proven to be a:
- fraud,
- problem gambler,
- shyster,
- liar,
- flamer,
- troll,
- baiter,
- attention whore
And worst of all, people lapped it up. His thread was finally locked after years of nonsensical and decreasingly sophisticated trolling.
I probably shouldn't have fed the troll then. 
Roby for sure.
