Certified Legendary Thread The Squiggle is back in 2023 (and other analytics)

[the harry]

It's worth noting that a really good model (or the betting line) gets the predicted margin wrong by 27 points per game on average. That's four and a half goals! And upsets are very common, occurring once every three or four games.

But correctly picking upsets is super difficult. At least in betting, when you tip an upset you get an oversized reward, which makes it worth doing more often. In a tipping comp, you get 1 tip no matter how unlikely it was. So you need to be smart or lucky, or both.

Of course, the person who wins the comp is always a contrarian, because that's what distinguishes them from the pack, but most contrarians will do worse than average.

So I'm tipping an upset every time the line is <13.5 points after this post.

 

[Billy ray]

So I'm tipping an upset every time the line is <13.5 points after this post.

where does one find the predicted margin each week to do this approach?

 

[WhiskeyTangoFoxtrot]

where does one find the predicted margin each week to do this approach?

Any sports bookmaker, i.e. TAB, Sportsbet, Bet365, CrownBet, etc.

 

[Suma Magic]

I’m not interested in overall, just winning a round or two (or more) in a 50-80 person comp.

I’ve decided on the following approach, based on differences between odds and tipping proportion.

This is an example using round 1 2019 data. Games in order eg tiges first. Proportion is % tipped divided by % chance per odds.



Interested in any feedback.

 

[Dr Tigris]

I’m not interested in overall, just winning a round or two (or more) in a 50-80 person comp.

I’ve decided on the following approach, based on differences between odds and tipping proportion.

This is an example using round 1 2019 data. Games in order eg tiges first. Proportion is % tipped divided by % chance per odds.

View attachment 633604

Interested in any feedback.

I just do it last second using Squiggle, the odds and my gut. I always end up top 10-20%. Only own the office tipping once when I was out of the country for the last 9 rounds and so did them all before I left

 

[DanWCE]

I’m not interested in overall, just winning a round or two (or more) in a 50-80 person comp.

I’ve decided on the following approach, based on differences between odds and tipping proportion.

This is an example using round 1 2019 data. Games in order eg tiges first. Proportion is % tipped divided by % chance per odds.

View attachment 633604

Interested in any feedback.

You're not in my tipping comp are you? It's a beauty for trying to pull off what you are..

 

[Final Siren]

I’m not interested in overall, just winning a round or two (or more) in a 50-80 person comp.

I’ve decided on the following approach, based on differences between odds and tipping proportion.

This is an example using round 1 2019 data. Games in order eg tiges first. Proportion is % tipped divided by % chance per odds.

View attachment 633604

Interested in any feedback.

I think that's a great way to do it. Whether the values you've chosen (e.g. "80%", "less than 10%") are optimal will depend on the makeup of your comp, so you could adjust as you go based on whether you would be winning the week if all your tips had come up -- if you'd have gotten 9 and the next closest person would only have gotten 6 or 7, you're tipping too many upsets; if several other people also would have gotten 8 or 9, you're not tipping enough upsets.

 

[iluvparis]

I normally use similar logic once I get behind in a tipping comp - tip 6 or 7 faves and look for 2-3 upsets each week - if you nail the upsets you shoot up the rankings really quick. But if you miss the upsets and a couple of the fave tips lose you end up with horrific weekly scores. You have to be pretty high risk to win the big comps though.

 

[Final Siren]

Hey! I'm on a podcast, talking about why Richmond fans are lunatics, BigFooty is a stew, and stats are awesome.

Find it here:

https://player.whooshkaa.com/episode?id=347991

It's a new podcast called "Chilling With Charlie," running interviews with various sports stats people: https://player.whooshkaa.com/shows/chilling-with-charlie

You can find it on the iPhone Podcast app too.

 

[Chief]

Hey! I'm on a podcast, talking about why Richmond fans are lunatics, BigFooty is a stew, and stats are awesome.

Find it here:

https://player.whooshkaa.com/episode?id=347991

It's a new podcast called "Chilling With Charlie," running interviews with various sports stats people: https://player.whooshkaa.com/shows/chilling-with-charlie

You can find it on the iPhone Podcast app too.

Tweeted.

 

[Dr Tigris]

Hey! I'm on a podcast, talking about why Richmond fans are lunatics, BigFooty is a stew, and stats are awesome.

Find it here:

https://player.whooshkaa.com/episode?id=347991

It's a new podcast called "Chilling With Charlie," running interviews with various sports stats people: https://player.whooshkaa.com/shows/chilling-with-charlie

You can find it on the iPhone Podcast app too.

Very good. You've got an interesting background. And the clubs could learn a fair bit from the data based approach. And old players / commentators could learn a lot

 

[Hobbes]

Now this thread has come back to life, I'm shifting my algorithm thread here. (Originally posted on Thursday morning, before the Richmond game).

Ratings from the end of last season

1. West Coast 26.4
2. Collingwood 22.7
3. Melbourne 19.4
4. Richmond 12.4
5. Geelong 11.7
6. GWS 9.9
7. Adelaide 8.5
8. Essendon 7.2
9. Hawthorn 2.9
10. North Melbourne 2.5
11. Western Bulldogs -1.9
12. Brisbane -2.2
13. Port Adelaide -2.3
14. Sydney -2.4
15. St Kilda -8.0
16. Fremantle -11.3
17. Gold Coast -16.2
18. Carlton -18.7

Carlton v Richmond +31
Collingwood +11v Geelong
Melbourne +27 v Port Adelaide
Adelaide +12 v Hawthorn
Western Bulldogs +7 v Sydney
Brisbane v West Coast +16
St Kilda +14 v Gold Coast
GWS +8 v Essendon
Fremantle v North Melbourne +5

I'll worry about season predictions once the football has actually started.

 

[rumply]

I’m not interested in overall, just winning a round or two (or more) in a 50-80 person comp.

I’ve decided on the following approach, based on differences between odds and tipping proportion.

This is an example using round 1 2019 data. Games in order eg tiges first. Proportion is % tipped divided by % chance per odds.

View attachment 633604

Interested in any feedback.

Interesting approach so I'm wondering where you get the % tipped figures from? There's the whotippedwhat website but those differ a fair bit from yours above.

 

[Suma Magic]

Interesting approach so I'm wondering where you get the % tipped figures from? There's the whotippedwhat website but those differ a fair bit from yours above.

Just from the espn footytips site. They are now out of date.

I picked the margin thurs night and got cats last night. Tipping all upsets from now on except north and crows.

 

[treefingers]

Very early days, but it's funny what a couple of r1 upsets can do to the forecasts:



I like this

 

[Billy ray]

Very early days, but it's funny what a couple of r1 upsets can do to the forecasts:

View attachment 640341

I like this

I wonder if the suns win by 100 tomorrow what it would show

 

[HTT]

Very early days, but it's funny what a couple of r1 upsets can do to the forecasts:

View attachment 640341

I like this

Reasonably comfortable away wins against highly rated opposition.
Plus it’s early in the year and the games had comfortable set shot wins too. It’s everything the squiggle loves.

 

[HTT]

Squiggle now has us finishing 5th, with a lower flagpole than Hawthorn and Port, but we knock off both of them on the way to the flag. I’ll take that.

 

[Pessimistic]

Its all confirming an even season. The fourth in a row which seems to be unprecedented

 

[Final Siren]

Squiggle is very reactive in the early rounds, so these big upsets are causing chaos!

 

[Cuzz09]

Just did a rapid fire prediction.

Lot a huge amount of thought and mainly going off gut feel as they pop up.

Eagles are going to have to win quite a few away games this year to make top 4.

That was a big loss last night

On SM-G925I using BigFooty.com mobile app

 

[BF Tiger]

View attachment 640630
Squiggle now has us finishing 5th, with a lower flagpole than Hawthorn and Port, but we knock off both of them on the way to the flag. I’ll take that.

You mean on the way to the flag we beat Geelong in their home final at the MCG? Delicious.

 

[Roobs321]

Love that reactive R1 squiggle, looks like some synchronised manoeuvre. The head-on collision between Brisbane and Melbourne is brutal!

 

[B4Bear]

It is going to blow a gasket after the Essendon debacle.

 

[holybishop]

I approve