Certified Legendary Thread Race for the flag, in squiggly lines

[Final Siren]

Round 2, 2015

Animated!

The squiggle loves a good thrashing, and we got plenty of them this round. It especially liked Sydney touching up Port, since this involved four things the squiggle rates:

1. Beating a good team
   
2. Winning by a lot
3. Winning interstate
4. Holding an opposition to a low score

Fremantle did almost the same thing, only less so.

Essendon's defence has been pretty amazing for the second week in a row, which is why they are sliding directly right. It's not unusual for a side to kick 70-odd points against the Hawks; Sydney did that in the Grand Final last year. But it is unusual to keep Hawthorn to the same kind of score. I think it's the first time the Hawks have scored in the 70s since 2013.

U-turns by West Coast and North, but a second successive week of bad falls for Geelong, Gold Coast, and Brisbane.

The Bulldogs jumped a lot on the squiggle predictor by edging to the top of the big pack of bottom-8 teams. There's still a fair gap to the top 10, though.

Squiggle predictor says:

Note: No Ninthmond jokes will be accepted.

All this and more in the interactive squiggle!

 

[Final Siren]

I think that is partly a result of a statistical quirk in the squiggle where it over-values keeping teams to a low score.

The squiggle does overvalue the situation where a team keeps its opposition to a really low score, like in the 20s. That's where the numbers start getting ridiculous. But aside from this, it's a feature, not a bug, as teams that can hold their opponents to low scores are genuinely good teams, and you should tip them.

So I think Sydney's current position is legit.

 

[Final Siren]

Final Siren are you able to post the Squiggle's Round 1 tips? I didn't catch them before the Tuesday update.

BTW - it would be awesome if we could some how look back at the tips for any round. If not, then I'm happy to manually log them to for the avid followers.

I don't have a way to go back and see tips for past rounds, sorry, no. If you want to collect them, that would be ace. Here is Round 1.

 

[The Swans Blog]

Final Siren the round 2 tips are still up I want to see how much your app is tipping us to thrash gonads wash by.

 

[Final Siren]

Final Siren the round 2 tips are still up I want to see how much your app is tipping us to thrash gonads wash by.

The tips stay up each week until Tuesday at noon, just so you can see what they were. You can peek ahead, though, using the Predictor -> Home & Away.

In Round 3, the Swans need to beat GWS 103-58 to hold their current squiggle position.

 

[Final Siren]

Squiggle has Adelaide finishing 4th, but has a 5th placed Port defeating them by 1 point in the semi. That would be entertaining.

Yeah, the squiggle still ranks Port ahead of Adelaide at the moment (just), but the Crows have two more wins in the bank with an easier fixture, so are tipped to finish the H&A season higher.

 

[bomberclifford]

Absolutely, I'd even pay to see that game @ AO

You can have my ticket because I don't think my heart would survive that.

 

[The Swans Blog]

Yah, I did check that Final Siren but is it normal that during a round if I drag a team even slightly it totally banjaxes the predictions? Does the prediction still take into account the game to be played, rather than the prediction at the end of the round? Eg. Hawthorn didn't play yet, but if I altered Freo, Swans or Hawks even the slightest, Swans drop 2 games and Hawks win the GF by 20+.

 

[Paul Kersey]

Keep it up.

 

[Roo Lives Again]

You can have my ticket because I don't think my heart would survive that.

AO would be rocking. Full house. Atmosphere would be electric. It'd be hard to take but an unmissable game.

 

[The Swans Blog]

You can have my ticket because I don't think my heart would survive that.

Heh heh heh I may take you up on that offer. What's the date for the game anyway?

 

[Final Siren]

Yah, I did check that Final Siren but is it normal that during a round if I drag a team even slightly it totally banjaxes the predictions? Does the prediction still take into account the game to be played, rather than the prediction at the end of the round? Eg. Hawthorn didn't play yet, but if I altered Freo, Swans or Hawks even the slightest, Swans drop 2 games and Hawks win the GF by 20+.

I'm not completely sure what you're doing... using the RECALCULATE button, right? Maybe it does get weird if you do that mid-round... I don't think I've ever actually tried. Seems OK now though.

You are right that it doesn't take much adjustment to create a wild change in the final prediction, since there are plenty of games between now and then. Especially if you're looking at a Grand Final involving teams from different states, since in that case, assuming they're fairly evenly matched, what will separate them will be home game advantage. This year, that means the nominal home team, regardless of where the game is actually played, as discussed earlier in the thread. For a Sydney vs Hawthorn Grand Final, the nominal home team is whoever finished on top of the ladder in the H&A season, and that's who will get the squiggle's "home game advantage" in the GF.

Bear in mind, though, the squiggle isn't designed to be a great GF tipper, and hasn't performed very well at that. It's good at tipping the regular season, and predicting a likely final ladder, because that requires juggling more data than human beings are comfortable with. But it's probably too simple and ignorant of too many factors to be better than human at tipping Grand Finals. Also, since there's only one GF per year, it's near-impossible to tell how much of a good tip was luck and how much was insight.

 

[bomberclifford]

AO would be rocking. Full house. Atmosphere would be electric. It'd be hard to take but an unmissable game.

True. Port win and my family wouldn't see me for at least a week, lose and it would time to remove my shoelaces and lock me in the spare room.

 

[poshman]

All this and more in the interactive squiggle![/QUOTE]

Love the squiggle.. just wondering why the predictor has Freo having lost a game at round five when they are tipped to win through. Apologies in advance if this has been explained before. Thanks!

 

[bomberclifford]

Heh heh heh I may take you up on that offer. What's the date for the game anyway?

Not sure. The squiggle doesn't say. #flawedsystem

 

[Pessimistic]

Yeah, the squiggle still ranks Port ahead of Adelaide at the moment (just), but the Crows have two more wins in the bank with an easier fixture, so are tipped to finish the H&A season higher.

If that were the scenario, finishing 2nd would be arguably a better run than 1st

 

[Final Siren]

Love the squiggle.. just wondering why the predictor has Freo having lost a game at round five when they are tipped to win through. Apologies in advance if this has been explained before. Thanks!

It's explained under INFO, because, yeah, it comes up a lot! I should make a little WELCOME TO SQUIGGLE badge for everyone who asks it.

See here:

Season Predictor

This is an estimate of how the ladder is most likely to turn out, on the assumption that the squiggle has correctly assessed the strength of every team and nobody gets much better or worse.

For the home & away season, it uses a probabilistic ladder. That means that instead of tipping individual games and tallying up the tips, it awards a win probability for each game. For example, if the squiggle thinks Hawthorn is 68% likely to beat Collingwood, it will award 0.68 wins to the Hawks and 0.32 to the Pies, increasing both team's tally of "probable wins" by less than 1.

This is because if a team plays 10 games with 60% likelihood of winning each game, we should expect them to win about 6/10—not, as we would get if we tipped each game and tallied up the tips, 10/10. We know that upsets will happen; we just don't know when. A probabilitistic ladder accounts for the likelihood that teams will sometimes unexpectedly win or lose, even though we doesn't know when.

This can look like a bug in the predictor, if you see that a team is predicted to win a match but it doesn't seem to be credited for it. For example, a team might be on "15 (14.7)" wins: that is, 14.7 "probable wins" rounded off to 15. (Rounding occurs so that teams are secondarily ranked by their percentage.) And then that team is tipped to win the following week, but it remains on 15 wins, now "15 (15.3)". What has happened is the number of probable wins hasn't risen by enough to be rounded to a higher number: it has earned 0.6 more probable wins but this still rounds off to 15. The predictor is saying it's still most likely this team will be on 15 wins, after accounting for the likelihood that some of its tips will be wrong.

 

[The Swans Blog]

I'm not completely sure what you're doing... using the RECALCULATE button, right? Maybe it does get weird if you do that mid-round... I don't think I've ever actually tried. Seems OK now though.

I was just shifting teams mid-round guessing their finishing positions of the round and hitting recalculate, but I'm guessing it still calculates the games to play. I could have moved the Swans 5+ across to the right and the Hawks would still have finished 1st and Freo 2nd before the Hawks played. Now of course the Swans perform far better.

You are right that it doesn't take much adjustment to create a wild change in the final prediction, since there are plenty of games between now and then. Especially if you're looking at a Grand Final involving teams from different states, since in that case, assuming they're fairly evenly matched, what will separate them will be home game advantage. This year, that means the nominal home team, regardless of where the game is actually played, as discussed earlier in the thread. For a Sydney vs Hawthorn Grand Final, the nominal home team is whoever finished on top of the ladder in the H&A season, and that's who will get the squiggle's "home game advantage" in the GF.

It was the same last season but couldn't explain the poor Sydney performance - nothing really could.

Bear in mind, though, the squiggle isn't designed to be a great GF tipper, and hasn't performed very well at that. It's good at tipping the regular season, and predicting a likely final ladder, because that requires juggling more data than human beings are comfortable with. But it's probably too simple and ignorant of too many factors to be better than human at tipping Grand Finals. Also, since there's only one GF per year, it's near-impossible to tell how much of a good tip was luck and how much was insight.

That's a good point, but I like that at the end of the year you go through your algorithms and give a side-by-side comparison and a reasoned tip for the winner.

 

[clogged]

There is a very good chance we may never see this poster again

Unfortunately not.

 

[The Swans Blog]

There is a very good chance we may never see this poster again

You wish..

 

[The Swans Blog]

Final Siren back to doing the PR vs SQ
https://docs.google.com/spreadsheets/d/1WYlVcE46usY2hs22BknuixuHShM3yNvOElvHOCFX-QE

Round 2:

H2H
PR: 8
SQ: 10

Tips:
PR: 12
SQ: 10

 

[TBellicious]

I love this. I sort of understand it, but i still love it.

Just a question for you Final Siren, does this give any indication on who it thinks will win the flag? Does it show anything other than the attack and defence of a team? I can understand it all but i just am having trouble interpreting it all into information.

 

[TBellicious]

Also, how come they started at different positions. I assume that those positions are where they are projected to start but wouldn't it be more logical that all teams start at the same position thus giving a better indication of how a team has gone throughout the year.

 

[The Swans Blog]

Also, how come they started at different positions. I assume that those positions are where they are projected to start but wouldn't it be more logical that all teams start at the same position thus giving a better indication of how a team has gone throughout the year.

Each team starts 50-50 for attack and defence, then using the finishing positions from the previous seasons as well as the results, he predicts their starting position for the next season.

How are the year's starting values calculated?

2015 starting positions are very similar to their end 2014 positions—the only difference is that 2013 data is no longer considered, so teams are modeled from the start of 2014 with each beginning on 50 ATTACK and 50 DEFENCE.

This means late-season 2014 results weigh quite heavily. For example, Collingwood had an injury-plagued end to 2014, and so is rated very low. Adelaide and West Coast, by contrast, finished the year with several solid performances, and so begin the year higher than you might expect.

 

[The Swans Blog]

Final Siren You should embed my Google Spreadsheet on the first page of this thread. I'll be updating it throughout the 2015 season.

This is what it looks like: