Certified Legendary Thread Race for the flag, in squiggly lines

[smasha]

Watching poo teams waiting for the big one.
#weareveryonesgrandfinal

 

[crazywildhorse]

West coast are going to make the top 4. They have weathered their poor early season form and will come home with the wind in their sails.

 

[Hobbes]

West coast are going to make the top 4. They have weathered their poor early season form and will come home with the wind in their sails.

Unlikely. Their form is good enough to make the top 4, but they're a match behind the pack and they have away matches against GWS and Adelaide to come, as well as a home match against Hawthorn they could easily lose. 14 wins wouldn't make the top 4, and 15 would be unlikely to. To be a real chance, they can afford only one loss from that run home, and I don't think that will happen.

 

[crazywildhorse]

Unlikely. Their form is good enough to make the top 4, but they're a match behind the pack and they have away matches against GWS and Adelaide to come, as well as a home match against Hawthorn they could easily lose. 14 wins wouldn't make the top 4, and 15 would be unlikely to. To be a real chance, they can afford only one loss from that run home, and I don't think that will happen.

They will beat Hawthorn and GWS, and may beat Adelaide as well. I still think 2016 GF will be a Haw vs WCE affair with West Coast putting in a much improved effort. Squiggle has had issues with Adelaide before, and despite being one of the teams that wouldn't surprise if they pulled it off, I still quite see it yet.

 

[Murray2503]

They will beat Hawthorn and GWS, and may beat Adelaide as well. I still think 2016 GF will be a Haw vs WCE affair with West Coast putting in a much improved effort. Squiggle has had issues with Adelaide before, and despite being one of the teams that wouldn't surprise if they pulled it off, I still quite see it yet.

Unlikely, is shaping up as a sydney v adelaide grandfinal in my opinion especially if crows beat cats round 18 as swans just did! They are the best attack v defence teams all year which normally play off in the grand final!

 

[Millky95]

I've gotta say, for a team whose deficiency was supposed to be their defence, and I thought they were suspect there, Sydney are doing incredibly well in that respect. With the Crows far above all others in attack, an Adelaide v Sydney GF would be a rip snorter.

We are only that far across because of two insanely wet games. We'd be much further left if those games weren't so wet. I suspect we'd be around 60 attack and 72 defense instead of 59 attack and 78 defense that we are now

 

[WhiskeyTangoFoxtrot]

Unlikely, is shaping up as a sydney v adelaide grandfinal in my opinion especially if crows beat cats round 18 as swans just did! They are the best attack v defence teams all year which normally play off in the grand final!

We're a long way out, but I think if that if that match-up did happen, and the game was anything like the game played earlier in the season between the 2 teams, most neutral supporters would be happy with that as a grand final.

It definitely had a finals-like intensity to it.

 

[chazwazza]

It's too close on the ladder and too far out from finals to suggest a top 4 and flag contenders. There's 8 clubs that have a realistic chance for a top 4 finish if not top 2. Only when this is settled will we get an idea of real contenders as the home finals will be crucial in such a tight year.

 

[EagleShore]

We are only that far across because of two insanely wet games. We'd be much further left if those games weren't so wet. I suspect we'd be around 60 attack and 72 defense instead of 59 attack and 78 defense that we are now

but you'd still not be weak defensively as many predicted. Rampe's season has been stellar.

 

[Millky95]

but you'd still not be weak defensively as many predicted. Rampe's season has been stellar.

And Grundy has been in AA form too. Rampe will make the team, Grundy the squad. Hopefully Aliir can hold his spot for the rest of the season too

 

[NikkiNoo]

True true but the rain helped. Squiggle must be an indoors character.
Interesting times ahead.

Lucky they picked Armfield to play otherwise it might have only been 24 points

 

[Gizza]

View attachment 264914 It defies logic that the top of the ladder team, six wins in a row, will lose more games than any of the other top 7View attachment 264891View attachment 264908

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.

 

[Final Siren]

Round 16, 2016

​

Animated!

Travelling interstate to hold your opponent to a low score is always a reliable way to get yourself squiggling, and that's what the Crows did after winning 107-47 over Carlton. If they weren't already, they're now clearly in the best position on the board.

Sydney had a great week, too, getting some valuable vertical movement from a convincing victory over Geelong at Kardinia. This actually makes the Swans the #1 ranked team on raw squiggle scores; that is, they're the hardest team to beat in the regular season. But they're also still in an area of the chart that historically produces many finalists and few premiers. So right now, you would rather be Adelaide.

West Coast did the business over North Melbourne, and Hawthorn over Port Adelaide, and that's the end of the list of top 8 teams that had good weeks.

GWS have the speed wobbles after their dizzying ascent this year, with one mediocre result and two poor ones from their last three games. North and Geelong were soundly beaten, and the Bulldogs had a dishonourable 10-pt victory over Richmond. Although I hope that was because the Tigers played like a real team for once, rather than because the Bulldogs aren't one.

Outside the finalists, Collingwood moved a lot, while Carlton and Fremantle sank. Seriously, Fremantle. When you can only score 55 against Melbourne, you have problems. I think Dees supporters will be with me on this one. At least Carlton were playing the squiggle-anointed premier.

And after letting through 22 goals from one of the league's weakest attacks, Brisbane's defence rating has become so poor, I almost had to adjust the chart axis. We have to go back to Carlton 2007 to find a team that leaked goals this badly, excluding expansion teams in their first year. They're even worse defensively than Melbourne 2013, who are my usual modern benchmark for badness.

Ladder Predictor! Jostling continues at the top, although there's more certainty than we've had for a while:

As we all know, the Ladder Predictor is built from probabilities, factoring in the likelihood of upsets, which is more realistic than assuming favourites always win. But as we get down to the pointy end of the season, it's worth looking at that version, too. I call it the DRAMATIC LADDER PREDICTOR, because it produces extreme results, like winless seasons:

Comparing the regular version to the DRAMATIC LADDER PREDICTOR, mostly what we see is that Sydney are in the box seat for the minor premiership. They finish on top either way: whether you weigh up the probabilities of upsets or whether you assume that favourites always win, like the DRAMATIC LADDER PREDICTOR. Adelaide also look good for top 2.

Geelong, meanwhile, nets 17 wins if they do what they're supposed to, but only 15 wins if we factor in the chance of upsets! And it's not even because Geelong are great at losing matches they're supposed to win. Like, dramatically great. It's just because they have a few games where they'll probably win but may not, like Adelaide and the Bulldogs (both at Kardinia).

Hawthorn are in a similar boat, with games they should win but might not against North and Collingwood.

And GWS are the only team more likely to unexpectedly pick up a win than drop one. That's because their expected wins are mostly gimme games (Brisbane, GC) while their expected losses are pretty winnable (North away, Port away, WC away).

Of course, it's all still very close. Hawthorn and Geelong are also, as it happens, being a little underrated by the regular ladder predictor, since they're both effectively predicted to win half a game more, but that's being rounded down, while Sydney are expected to win almost half a game less, which is being rounded up. Bottom line, it'll come down to which teams can perform when it counts, and a single result -- like Sydney vs Hawthorn this Thursday! -- can change everything.

Flagpole! I've gotta say, Flagpole has done well so far, I reckon. There were times when I looked at it this year and thought, "Adelaide, really?" Or that Hawthorn should have fallen further faster, and Geelong and GWS risen faster. But here at Round 16, the same top 3 it's had all year of Adelaide, Hawthorn and West Coast looks pretty good. And it's been a tough year to tip, with 8 genuine contenders.

Eagle eyes already spotted that West Coast just baaaaaaarely squeaks over Hawthorn here. But more significant is that 4th spot changing again, this time going to Sydney.

Live squiggling! Tell your friends: It's a probabilistic ladder.

 

[Final Siren]

I've gotta say, for a team whose deficiency was supposed to be their defence, and I thought they were suspect there, Sydney are doing incredibly well in that respect. With the Crows far above all others in attack, an Adelaide v Sydney GF would be a rip snorter.

We had a GF featuring the year's attack specialist vs the year's defence specialist in 2013, and I was all excited about it, then it turned out to be a fizzer. Hopefully Crows vs Swans would be better than that.

 

[WhiskeyTangoFoxtrot]

Final Siren, how does the model deal with the fact that the final number of wins predicted may be more than the actual number of wins possible due to rounding?

For example, the model may predict that 4 teams finish with 15.51 wins (which rounds to 16 wins) and another team finishes with 17.96 wins (which rounds to 18 wins), a total of 82 wins amongst those 5 teams, when in reality there should only be 80 possible wins amongst those 5 teams.

 

[chazwazza]

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.

 

[DrEvil_]

I've gotta say, for a team whose deficiency was supposed to be their defence, and I thought they were suspect there, Sydney are doing incredibly well in that respect. With the Crows far above all others in attack, an Adelaide v Sydney GF would be a rip snorter.

Sydney played quite a few games in terrible wet conditions, inflating their defense and underestimating their offence.

 

[Gralin]

Sydney played quite a few games in terrible wet conditions, inflating their defense and underestimating their offence.

True but by how much is the question

 

[Final Siren]

Final Siren, how does the model deal with the fact that the final number of wins predicted may be more than the actual number of wins possible due to rounding?

For example, the model may predict that 4 teams finish with 15.51 wins (which rounds to 16 wins) and another team finishes with 17.96 wins (which rounds to 18 wins), a total of 82 wins amongst those 5 teams, when in reality there should only be 80 possible wins amongst those 5 teams.

Right! It deals with this by blatantly ignoring it.

How you can actually solve that problem is you run simulations many thousands of times and use the ladder that results most often. But this is too slow for a web tool that updates in near-real time, still imperfect, because you can get teams that do well in lots of ladders beside the most common one, and in practice produces very similar results to the quick squiggle one.

There might also be some super mathy method with path following but that's beyond my pay grade.

 

[GreyCrow]

Looks like the Crows are aiming for the carpark between Brisbane 2001,02 Geelong 2007 and Hawthorn 2014

 

[1990crow]

Looks like the Crows are aiming for the carpark between Brisbane 2001,02 Geelong 2007 and Hawthorn 2014

Some average sides between them.

 

[WhiskeyTangoFoxtrot]

Right! It deals with this by blatantly ignoring it.

How you can actually solve that problem is you run simulations many thousands of times and use the ladder that results most often. But this is too slow for a web tool that updates in near-real time, still imperfect, because you can get teams that do well in lots of ladders beside the most common one, and in practice produces very similar results to the quick squiggle one.

There might also be some super mathy method with path following but that's beyond my pay grade.

For the record, I think close enough is good enough. I was just wondering how it dealt with that sort of problem (which doesn't come up too much anyway). Turns out it doesn't!

 

[DrEvil_]

Watching poo teams waiting for the big one.
#weareveryonesgrandfinal

Where are you wearing it?

 

[cryptor]

Right! It deals with this by blatantly ignoring it.

How you can actually solve that problem is you run simulations many thousands of times and use the ladder that results most often. But this is too slow for a web tool that updates in near-real time, still imperfect, because you can get teams that do well in lots of ladders beside the most common one, and in practice produces very similar results to the quick squiggle one.

There might also be some super mathy method with path following but that's beyond my pay grade.

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.

 

[Murray2503]

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.

North have copped injuries at the worst time but i dont see them as a premiership threat even with their best side as they will need to beat either wce, sydney, adelaide, hawks away in the finals. I think they are no hope without wells!