Analysis Using Data to Identify Midfield Targets

[Lazzas]

Thankyou Dylan,

Fantastic work as always. Garcia and Windhager rate highly, I hope the Eagles are chasing a couple of player's ready to step up.

 

[FKASC]

Will Phillips pops up a bit in your tables. He’s a bit of a meme for Norf’s shit drafting but might be worth looking at if cheap? Surely he’d be keen for some more consistent opportunity? Or just too spudly

 

[Ambrose Burnside]

Love this kinda shit.

Apologies if it's been done, but how did the model rate the ones we have picked up over the last few years, like Hunt, JJ, Brockman etc?

 

[Mattman93]

Will Phillips pops up a bit in your tables. He’s a bit of a meme for Norf’s shit drafting but might be worth looking at if cheap? Surely he’d be keen for some more consistent opportunity? Or just too spudly

I'd hoped so, but just signed a one year deal.

 

[Yeahnahmate]

Dylan82 for PM

 

[DanWCE]

Dylan82 for PM

Procuring Midfielders?

 

[OldBlueEagle]

Great job Dylan, really interesting

 

[Dylan82]

Will Phillips pops up a bit in your tables. He’s a bit of a meme for Norf’s shit drafting but might be worth looking at if cheap? Surely he’d be keen for some more consistent opportunity? Or just too spudly

You're correct, he [W.Phillips] does come up a lot and I were to have an 11th candidate, it would probably have been him.

Main reason for his exclusion was the lower output at VFL level this year in combination with his smaller size and the likelihood that North won't consider on any terms trading a No.3 pick that hasn't even played 50 games.


Similarly, J.Burgoyne didn't figure despite some strong showings among models as there is zero chance he will move away from Port.

 

[Dylan82]

Hi Dylan

Looks like a lot of work and lots information to produce some pretty ordinary outcomes.

The goal is an exercise in trying to highlight talent prior to it breaking out as a means of possibly gaining it cheaply.

A player with unrealised potential could well be expected to look rather ordinary until the right environment/opportunity arises.


At the very least, it provides some suggestive evidence to corroborate targets the club is being linked to may be able to contribute and improve beyond what we have seen from them so far to date.

 

[Dylan82]

Love this kinda shit.

Apologies if it's been done, but how did the model rate the ones we have picked up over the last few years, like Hunt, JJ, Brockman etc?

Looking back, none of those you mention have figured on any of the models.


S.Petrevski-Seton was flagged under the opportunity model in 2019, but he did not join the club until the 2022 season, by which stage it had dropped him entirely.

The opportunity model is built around how Yeo developed as a late bloomer compared to many in his draft cohort, so naturally he appears there in 2017 prior to his breakout as a midfielder in 2018.

Before that you have to go to J.Redden who had multiple years (2011, 2012, 2013 & 2014) at the Lions with significant figures (including +5.6 in 2012 as a 21 year old) for the AI Predicted Improvement model.

S.Wellingham also had multiple seasons (2010, 2011) where he figured below -1.0 (-2.2 in 2010) for the AI Predicted Improvement model.


Considering the magnitude of the trade, the closest T.Kelly has come to figuring on any of the models is -0.3 for the AI Predicted Improvement model in 2018.

 

[RookiePick]

S.Wellingham also had multiple seasons (2010, 2011) where he figured below -1.0 (-2.2 in 2010) for the AI Predicted Improvement model.

You can't improve perfection.

 

[dinnaz]

Great write up.

However how much more advantage does a deep dive give us over the old fashioned stats from the newspaper:
Disposals
Tackles
Goals
Clearances
Clangers
etc

And the new stats:
Disposal Efficiency
Metres Gained
Contested possessions
Etc

I'm sceptical looking for trends as development is often non linear and can have many more factors at play.

Namely being played out of position/role compared to a similar prospect