2014 Brownlow Medal

Remove this Banner Ad

Log in to remove this ad.

I've built a reasonably sophisticated logistic regression model based on stats and some other factors I will keep to myself, which has performed admirably in backtesting. I'm happy to post the full output here but I can't for the life of me figure out how to attach a spreadsheet. I thought it may be of interest because it's a bit of a departure from most of the conversation I've seen around here. In particular has Boak beating Gray, Beams beating Pendles and Franklin tying JPK. I think it's main value lies in awarding votes to players in losing teams, which seems to be an area a lot of us struggle with.
 
I've built a reasonably sophisticated logistic regression model based on stats and some other factors I will keep to myself, which has performed admirably in backtesting. I'm happy to post the full output here but I can't for the life of me figure out how to attach a spreadsheet. I thought it may be of interest because it's a bit of a departure from most of the conversation I've seen around here. In particular has Boak beating Gray, Beams beating Pendles and Franklin tying JPK. I think it's main value lies in awarding votes to players in losing teams, which seems to be an area a lot of us struggle with.

Who's your top 5?
 

(Log in to remove this ad.)

Who's your top 5?
Ablett 26
Fyfe 25
Priddis 23
Selwood 22
Rockliff 21

And three more to give an eligible top 5...
Johnson 21
Cotchin 20
Beams 20

Remember this is no input from me, purely statistics based with some media input, so please keep any flames to yourself. Looking at it subjectively, Rockliff is probably the furthest out as Ablett picking up votes in big losing sides has influenced the model fit somewhat. I'm actually considering a refit with Gary receiving his own variable.
 
By looking at the photo I'm guessing the last 3 "specials" will be Selwood, Ablett and ???

Heppell and Priddis were easy to guess with the hair haha. The one to the right of the logo looks like the bald Gary to me.
 
I've built a reasonably sophisticated logistic regression model based on stats and some other factors I will keep to myself, which has performed admirably in backtesting. I'm happy to post the full output here but I can't for the life of me figure out how to attach a spreadsheet. I thought it may be of interest because it's a bit of a departure from most of the conversation I've seen around here. In particular has Boak beating Gray, Beams beating Pendles and Franklin tying JPK. I think it's main value lies in awarding votes to players in losing teams, which seems to be an area a lot of us struggle with.
How does cotchin look in your model.... Looks good in mine
 

Remove this Banner Ad

2014 Brownlow Medal

Remove this Banner Ad

Back
Top