Models Leaderboard After Home & Away

2017 has been a rough year for tipping, even tougher than 2016, which itself was a wake-up call after the lovely, predictable Hawthorn dynasty.

Sadly, none of the models managed to outperform the bookies in tipping winners head-to-head during the Home & Away season, nor in terms of probabilistic Bits.

Tips Bits MAE Correct
Punters

4 8 4 7 7 6 3 5 5 7 4 4 3 7 5 5 7 6 6 9 8 7 5

132 22.1970 66.7%
PlusSixOne

5 8 4 7 6 6 3 5 6 7 3 3 2 8 5 7 7 5 6 5 8 7 6

129 17.9055 28.89 65.2%

5 9 4 6 7 6 3 5 4 7 4 3 2 6 5 5 6 7 7 8 6 6 7

128 17.5804 64.6%

3 8 5 6 7 7 4 5 4 7 4 1 3 6 4 6 8 7 7 8 6 6 6

128 16.3202 28.87 64.6%

5 7 5 6 6 6 4 5 7 7 3 2 3 7 3 5 7 5 7 7 7 7 7

128 15.4157 30.11 64.6%
Aggregate

5 8 4 6 7 6 3 5 5 7 3 3 2 6 4 5 7 6 7 8 7 6 5

125 20.1305 28.92 63.1%

5 7 5 6 6 6 3 5 5 7 3 3 3 6 4 6 7 6 6 7 7 6 6

125 16.5200 29.03 63.1%

4 7 4 6 7 6 3 5 5 7 4 3 2 6 3 5 6 6 7 8 8 6 5

123 18.6520 28.73 62.1%

5 8 5 5 5 6 3 5 5 8 4 2 2 7 3 6 7 6 6 7 6 5 6

122 17.0455 29.79 61.6%

A particularly rough year for Squiggle’s in-house tipping algorithm, which came in dead last and didn’t deserve to be any higher. But very credible performances elsewhere, especially from Plus Six One (for the second year in a row), and it’s worth noting The Arc‘s chart-topping Mean Average Error in a year when the line-ball games didn’t fall their way.

I was curious this year to see what kind of performance the Aggregate would have, where it represents a simple average of everyone else’s tips. Would there be some kind of wisdom of the crowds effect, where it could outperform most of the individual models that provided its inputs? Well, kind of: It landed mid-table in terms of tips and MAE, but with more Bits than all models. So there could be something there.

Apologies for finals tipping being a little shaky at the moment: Apparently everyone (including Squiggle) uses different formatting for posting their finals tips, and squiggle-bot needs to learn how to parse it.

Tip Leaderboard after Round 17

The Tip Leaderboard has developed a bit of spread over the last few weeks! PlusSixOne has pulled away to a whopping four tips lead over the next best model, and it’s leading in Bits as well:

Tips Bits
PlusSixOne

5 8 4 7 6 6 3 5 6 7 3 3 2 8 5 7 7

92 9.5796
Punters

4 8 4 7 7 6 3 5 5 7 4 4 3 7 5 5 7

91 8.0603

5 7 5 6 6 6 4 5 7 7 3 2 3 7 3 5 7

88 3.6086

3 8 5 6 7 7 4 5 4 7 4 1 3 6 4 6 8

88 2.7082

5 7 5 6 6 6 3 5 5 7 3 3 3 6 4 6 7

87 4.6444

5 9 4 6 7 6 3 5 4 7 4 3 2 6 5 5 6

87 4.0108
Aggregate

5 8 4 6 7 6 3 5 5 7 3 3 2 6 4 5 7

86 7.3906

5 8 5 5 5 6 3 5 5 8 4 2 2 7 3 6 7

86 4.1921

4 7 4 6 7 6 3 5 5 7 4 3 2 6 3 5 6

83 6.1518

“Bits” from Monash University Probabilistic Footy Tipping rewards tipsters for saying a win was likely and punishes them for saying it was unlikely.

New to Squiggle is Graft Ratings! Head on over for beautiful tips and projections of all kinds. Graft is having what appears to be a typical model year so far, sitting solidly in the 86-88 tip bracket.

And The Arc can be justifiably feeling a little screwed, with the worst tip numbers despite very respectable Bits and MAE (stats available on the main leaderboard page).

Collingwood under Buckley

As part of its weekly wrap-up, Footy Maths has a neat graphic depiction of the slide of the Pies under Nathan Buckley, and a measured take on what it means:

Both arrays of data show a decline since 2012. Collingwood has yet to become a sub-average rated team by the FMI measures, which is a limited positive.

Also, when compared to the Essendon journey since Sheedy, the decline has not been dramatic enough to have been sub-par as the Dons were three times before sanctions were levied. But also, after each dive there was a rally and development. The above chart has much less rallying, and more flatlining.

Source: FOOTYMATHS INSTITUTE: 16. TRANSITION RENDITION

The Bookies Bite Back

Punters have known best lately, with the favorites returning 7 tips in each of the last two rounds.

But Figuring Footy is still claiming first place off the back of a great season opening!

Tips Bits

5 9 4 6 7

31 4.1117
Punters

4 8 4 7 7

30 5.6630
PlusSixOne

5 8 4 7 6

30 4.6239
Aggregate

5 8 4 6 7

30 4.0893

5 7 5 6 6

29 1.0637

4 7 4 6 7

28 3.8395

3 8 4 6 7

28 3.3945

5 8 5 5 5

28 1.5787

Disaster for Squiggle, which went out on a limb on Round 5 tips for North Melbourne over Fremantle, and the Demons over the Tigers, and came up short on both.

You can now track these numbers in the new auto-updating Tip Leaderboard.

 

Round 1 Preview Roundup

A few musings on the shape of the season that’s about to begin:

  • Figuring Footy (warning: contains Eddie McGuire pic):

    2017 promises to be one of the most exciting seasons of footy in recent memory. There are probably 16 clubs with legitimate aspirations of, come September, giving the finals a big shake up the way the Dogs did last year.

    Source: My Focus for 2017 and Some Round 1 Tips – Figuring Footy

     

  • Matter of Stats gets wonky about Round 1 predictions and where they come from:

    This week we see that Home Sweet Home has the highest Disagreement Index (readers from last year will know that this is completely unsurprising), and that disagreement levels are generally high as the Tipsters are split 5-4 in three of the games, and 7-2 in two more.

    Source: 2017 – Round 1 : Is This Thing On?

     

  • FootyMaths Institute welcomes you to the new season with some old wisdom:

    Hey guess what? That new recruit you think is really good… yeah well this model (and other ELO models) don’t see things like that. Principle is every team has a 36 player list. One player falls, another rises to take their place. And really, everyone thought Adelaide would suffer without ‘Danger’ and the finished as defeated semi finalists.

    And there are fresh articles on Expected Wins and Expected Finish for each team as well.