Certified Legendary Thread Race for the flag, in squiggly lines

[lamby]

Sydney Squiggling into very familiar territory. I wish I had as much faith in Adelaide as The Squiggle does, on every ladder predictor I've done I don't have Adelaide making the top 2 or hosting a prelim and with that it's hard to see them getting to a GF.

Crows are also the only team who are yet to play Brisbane or Bombers!

Here are the Crows’s opposition scores in the 4th quarter:

Blues: 1.3
Demons: 1.3
Roos: 1.1
WCE: 0.0
Saints: 1.4

That is 5 games, , and a combined score of 4.11 for the last quarter. Two of those teams were against top 8 teams. 4 goals in 5 quarters of footy is some impressive defending. None of those games were against bottom 4 teams either.

 

[Hobbes]

You could deal with it using some logic algorithm which compares the total predicted wins with the total wins possible. Then if there's a difference (say +2 wins) then it would go through and find the teams with the lowest value of wins that are being rounded up (say 13.51 and 16.53) and then instead of rounding up you round down.

Well, some losses might be rounded up to a win. If there are eight wins to be accounted for, you could just take the eight largest partial credits.

 

[Final Siren]

You could deal with it using some logic algorithm which compares the total predicted wins with the total wins possible. Then if there's a difference (say +2 wins) then it would go through and find the teams with the lowest value of wins that are being rounded up (say 13.51 and 16.53) and then instead of rounding up you round down.

You could, but that means you'd care more about ensuring that the total number of wins add up rather than prediction accuracy.

Also it would kind of be the worst of both worlds, in that you still wouldn't have a guaranteed possible ladder, depending on who plays whom.

Imagine a situation where Hawthorn can only finish 1st or 5th, with equal chance of each. The Squiggle Ladder Predictor will put them in 3rd, because it cares most about getting each team as close to the likely outcome as possible. They're not going to actually finish 3rd but it's not going to be far off. A predictor that only produced possible ladders would choose either 1st or 5th and be exactly right or far off.

There are good sites out there that produce ladder predictors based on 10,000+ simulations. I'm not sure if they always produce guaranteed possible ladders but I would think so.

 

[Hobbes]

You could, but that means you'd care more about ensuring that the total number of wins add up rather than prediction accuracy.

Also it would kind of be the worst of both worlds, in that you still wouldn't have a guaranteed possible ladder, depending on who plays whom.

Imagine a situation where Hawthorn can only finish 1st or 5th, with equal chance of each. The Squiggle Ladder Predictor will put them in 3rd, because it cares most about getting each team as close to the likely outcome as possible. They're not going to actually finish 3rd but it's not going to be far off. A predictor that only produced possible ladders would choose either 1st or 5th and be exactly right or far off.

There are good sites out there that produce ladder predictors based on 10,000+ simulations. I'm not sure if they always produce guaranteed possible ladders but I would think so.

Except, I don't think that your ladder can do this. There's no team whose average result is first.

More likely, you have to shoehorn teams with an average final placing of 2.5, 2.9, 3.4, 4.0, 5.0, 5.8, 6.4, 7.0 into 1st-8th place.

 

[Simon_Nesbit]

Squiggle ranking the ladder for future rounds based on whole number rounding of wins - instead of rounding, perhaps ranked by part wins?

 

[GreyCrow]

I cant fathom why people want a ''prediction'' to be accurate. Its the best it is.

 

[Millky95]

Squiggle ranking from start of the year to now:

Adelaide 8 -> 2
Brisbane 15 -> 17
Carlton 18 -> 16
Collingwood 11 -> 10
Essendon 17 -> 18
Fremantle 9 -> 13
Geelong 10 -> 4
Gold Coast 16 -> 15
GWS 12 -> 8
Hawthorn 1 -> 5
Melbourne 13 -> 11
North Melbourne 5 -> 6
Port Adelaide 6 -> 9
Richmond 3 -> 12
St.Kilda 14 -> 14
Sydney 4 -> 1
West Coast 2 -> 3
Western Bulldogs 7 -> 7

 

[Pessimistic]

North were top of the ladder with 9 wins in a row and now have lost 5 of the last 6. Doesn't defy logic at all.

Hawthorn are yet to play Collingwood and Carlton, have three losses and 120%
North are yet to play Port Adelaide and GWS, have five losses and 112%

This shows they are hardly a top 8 benchmark any more, but even on the 'probability' score squiggle predicts 2.52 more losses for Hawthorn and 2.68 more losses for North, who have a considerably harder draw and are not travelling as well

Adelaide have the easiest remaining draw, have similar form to the Hawks, squiggle is predicting 1.6 more losses so I guess its finely balanced. Sydney similar

 

[GreyCrow]

Squiggle ranking from start of the year to now:

Adelaide 8 -> 2
Brisbane 15 -> 17
Carlton 18 -> 16
Collingwood 11 -> 10
Essendon 17 -> 18
Fremantle 9 -> 13
Geelong 10 -> 4
Gold Coast 16 -> 15
GWS 12 -> 8
Hawthorn 1 -> 5
Melbourne 13 -> 11
North Melbourne 5 -> 6
Port Adelaide 6 -> 9
Richmond 3 -> 12
St.Kilda 14 -> 14
Sydney 4 -> 1
West Coast 2 -> 3
Western Bulldogs 7 -> 7

11 within 3 ( I like that margin ) and another 4 within 4. Tipped 5 of the top 8

 

[Final Siren]

Speaking of ladder simulations! I finally added Squiggle Doors for the current season, so you can replay the year and see how things would look with random luck added.

Live Squiggle -> Doors

Why do this? Because luck really is a big part of football; it's not that good teams always win close games. Fremantle is the starkest recent example: The Dockers won an amazing number of close games last year while Hawthorn and West Coast lost them all, resulting in perhaps the luckiest minor premiership ever. A huge part of the reason for their fall in 2016 is that they weren't so great to begin with.

So by randomizing results a bit, you can see what the ladder would look like if the swings and roundabouts had swung and revolved the other way. By doing it multiple times, you get a sense for how lucky or unlucky some teams' position on the real ladder is.

Hawthorn are #1 on the ladder today, but with their surfeit of close wins, if only a few things had gone differently, they'd be 6th or 7th. As much as the ladder undersold them last year, it's flattering them right now.

Sydney, Adelaide, Geelong, and GWS all tend to gain from a little luck, because they've recorded close losses.

And Port could be within striking distance of the 8, threatening North Melbourne's spot.

In a season this even, a little luck goes a long way!

 

[Dannnnnnnnnn]

First random click with Doors:

https://maxbarry.com/squiggle/?door=3984407462&luck=43&year=2016&rewritehistory=1

Brisbane beat Hawthorn 170-81 in the grand final. Seems legit!

 

[Final Siren]

First random click with Doors:

https://maxbarry.com/squiggle/?door=3984407462&luck=43&year=2016&rewritehistory=1

Brisbane beat Hawthorn 170-81 in the grand final. Seems legit!

That is genuinely flukey even with 43% luck.

But obviously the more luck you add, the more random the results. Even 5% luck produces some big differences this year. I saw a ladder with Hawthorn out of the 8.

 

[hippyhippy]

Squiggle ranking from start of the year to now:

Adelaide 8 -> 2
Brisbane 15 -> 17
Carlton 18 -> 16
Collingwood 11 -> 10
Essendon 17 -> 18
Fremantle 9 -> 13
Geelong 10 -> 4
Gold Coast 16 -> 15
GWS 12 -> 8
Hawthorn 1 -> 5
Melbourne 13 -> 11
North Melbourne 5 -> 6
Port Adelaide 6 -> 9
Richmond 3 -> 12
St.Kilda 14 -> 14
Sydney 4 -> 1
West Coast 2 -> 3
Western Bulldogs 7 -> 7

Not even squiggle can understand Richmonds shitness. Every other team is around the mark

 

[KissKiss]

First random click with Doors:

https://maxbarry.com/squiggle/?door=3984407462&luck=43&year=2016&rewritehistory=1

Brisbane beat Hawthorn 170-81 in the grand final. Seems legit!

Especially after Hawks 172-85 Brisbane in week 1 of the finals. I like this scenario a lot!!!

 

[cryptor]

I like this one where the Hawks continue their pattern of winning their n'th flag in this run from finishing in the n'th position on the ladder...

https://maxbarry.com/squiggle/?door=660722704&luck=1&year=2016&rewritehistory=0

With only 1% luck it might not even be an alternative reality in a couple months

Or maybe we just do it from 5th instead?

https://maxbarry.com/squiggle/?door=3575756603&luck=1&year=2016&rewritehistory=0

 

[chazwazza]

Let's just hope that no one couples the Squiggle Doors and The Hadron Collider up together.

 

[RookiePick]

https://maxbarry.com/squiggle/?door=3399576023&luck=5&year=2016&rewritehistory=1

I'll take it!

 

[freo_leo29]

https://maxbarry.com/squiggle/?door=4087147358&luck=98&year=2016&rewritehistory=1#
second click and freo are absolutely destroying teams each week in the finals, 98% luck is possible next year... right??.. righttt?

 

[Here2tellyouwhy]

At the pre determined 5% luck the Hawks went to 8th and the GF played out like this...



Dangerfield would have to run over plenty of black cats you would think.

 

[radiojake]

Surely there is a Danderfield shank from 40 metres out after the siren there

 

[Here2tellyouwhy]

Surely there is a Danderfield shank from 40 metres out after the siren there

He hits a Crows fan in bay13 on the tit.

 

[Final Siren]

https://maxbarry.com/squiggle/?door=4087147358&luck=98&year=2016&rewritehistory=1#
second click and freo are absolutely destroying teams each week in the finals, 98% luck is possible next year... right??.. righttt?

I think you may have had 98% luck last year.

 

[Suma Magic]

Speaking of ladder simulations! I finally added Squiggle Doors for the current season, so you can replay the year and see how things would look with random luck added.

Live Squiggle -> Doors

Why do this? Because luck really is a big part of football; it's not that good teams always win close games. Fremantle is the starkest recent example: The Dockers won an amazing number of close games last year while Hawthorn and West Coast lost them all, resulting in perhaps the luckiest minor premiership ever. A huge part of the reason for their fall in 2016 is that they weren't so great to begin with.

So by randomizing results a bit, you can see what the ladder would look like if the swings and roundabouts had swung and revolved the other way. By doing it multiple times, you get a sense for how lucky or unlucky some teams' position on the real ladder is.

Hawthorn are #1 on the ladder today, but with their surfeit of close wins, if only a few things had gone differently, they'd be 6th or 7th. As much as the ladder undersold them last year, it's flattering them right now.

Sydney, Adelaide, Geelong, and GWS all tend to gain from a little luck, because they've recorded close losses.

And Port could be within striking distance of the 8, threatening North Melbourne's spot.

In a season this even, a little luck goes a long way!

How does the luck work again?

For example, if we increase "luck", does it mean that hawthorn are equally likely to win their close games by a larger margin than a smaller margin/lose?

Is the real season considered luck-free as the starting point? Sorry, I'm a bit confused. I don't understand how results that have occurred can be considered to be due to luck.

 

[Final Siren]

How does the luck work again?

For example, if we increase "luck", does it mean that hawthorn are equally likely to win their close games by a larger margin than a smaller margin/lose?

Is the real season considered luck-free as the starting point? Sorry, I'm a bit confused. I don't understand how results that have occurred can be considered to be due to luck.

Real results are considered to have 0 luck. As you increase luck, scores have an increasing amount of randomness applied to them.

This is completely unbiased but it's bad for teams that have won many close matches, like Hawthorn this year, because the randomness tends to turn some of those into losses.

A normal team, which wins about as many close ones as it loses, isn't reliably affected one way or the other.

I like to use it to see how knife-edge some situations are. For example this year, it takes quite a lot of luck to budge Richmond out of around 13th spot. We've won a game by 1 point and lost one by 1 point, have another close-ish win and a close-ish loss, and the rest are fairly decisive. Also, none of the teams around us tend to change much. So it's not like those years where I can bitch that the goal umpire in Perth cost us 4th spot.

But some teams, including the entire top 8, could be in very different positions with only a teensy bit of luck. And I think that's helpful to know, since otherwise you look at the ladder and naturally think that 1st is a lot better than 7th.

In a sense it's all a pointless exercise in "woulda shoulda coulda." But as a Richmond supporter that's about all I've had for 30 years so I'm writing web tools for it.

 

[Suma Magic]

Real results are considered to have 0 luck. As you increase luck, scores have an increasing amount of randomness applied to them.

This is completely unbiased but it's bad for teams that have won many close matches, like Hawthorn this year, because the randomness tends to turn some of those into losses.

A normal team, which wins about as many close ones as it loses, isn't reliably affected one way or the other.

I like to use it to see how knife-edge some situations are. For example this year, it takes quite a lot of luck to budge Richmond out of around 13th spot. We've won a game by 1 point and lost one by 1 point, have another close-ish win and a close-ish loss, and the rest are fairly decisive. Also, none of the teams around us tend to change much. So it's not like those years where I can bitch that the goal umpire in Perth cost us 4th spot.

But some teams, including the entire top 8, could be in very different positions with only a teensy bit of luck. And I think that's helpful to know, since otherwise you look at the ladder and naturally think that 1st is a lot better than 7th.

In a sense it's all a pointless exercise in "woulda shoulda coulda." But as a Richmond supporter that's about all I've had for 30 years so I'm writing web tools for it.

That's an excellent explanation. Cheers.