Analysis Champion Data Elite.

[Milanista28]

Damn straight. Hickey over Gawn every day of the week

Have Hickey in my keeper league, love him

 

[Kane McGoodwin]

No Walsh = fail list LOL

Play him on a half back flank & he will oust either Williams or Saad.

 

[Kane McGoodwin]

Port must be favourites for the flag after recruiting all time AFL hall of famer J Finlayson.

Just as well as apparently their B&F Brownlow medalist is not elite.

 

[Gysberts2Bate]

The Western with five elite midfielders, I’d hate to come up against them in a pressure situation trying to win centre clearance to get yourself back into the game.

 

[owen87]

The Western with five elite midfielders, I’d hate to come up against them in a pressure situation trying to win centre clearance to get yourself back into the game.

Shows the value of team and structure I guess, though Oliver & Petracca weren't missed on the list anyway.

 

[Scorpus]

Champion Data need to consult with the clubs over what statistics they value and completely re-do their subjective weighting system.

What the coaches value in a player is one of the most important metrics. So having B&F winners and 5 of the top 10 in the AFLCA award miss the "elite" list is a huge flaw, especially when some of those players are from finals teams.

Aside from that these are some of the clear flaws in their system that I've seen over the years:

- Not adequately rating defensive capabilities, particularly from players who are "lockdown" defenders, not offensive defenders. The ability to stop the opposition and stop goals is clearly valued by the league.

- Not rating uncontested possessions. I can't quite remember where but someone on here found that Champion Data effectively give zero worth to uncontested possessions, when surely having an uncontested possession is better than having no possession at all.

- Giving large penalties for kicking behinds or similar mistakes. For example in this game from Jake Riccardi, Champion Data thought that dropping a mark and having an ineffective handball was worth a greater penalty than the gain from kicking a goal, from a key position forward. That's a flawed weighting system

- Weighting hitouts and hitouts to advantage too highly. Rucks in general are positioned too high in their ratings when there is only a limited correlation between winning these stats and winning games. But then at times dominant around the ground play (eg Max Gawn

11

Max Gawn

Age

32

Ht

209cm

Wt

110kg

Pos.

Ruck

Career

Season

Last 5

D

14.1

4star

K

7.4

3star

HB

6.7

4star

CL

2.8

4star

HO

35.3

5star

D

18.5

4star

K

11.5

4star

HB

7.0

4star

CL

3.5

4star

HO

29.2

5star

D

8.4

3star

K

3.6

2star

HB

4.8

4star

CL

1.0

3star

HO

15.6

5star

) is not enough to become "elite"

- Incorrectly assigning positions to players they don't play in. For example at the Crows, Chayce Jones

1

Chayce Jones

Age

24

Ht

182cm

Wt

84kg

Pos.

D/M

Career

Season

Last 5

D

10.0

3star

K

5.4

2star

HB

4.6

3star

M

1.7

2star

T

1.9

4star

MG

178.6

3star

D

8.9

2star

K

4.9

2star

HB

4.0

3star

M

1.4

2star

T

1.6

3star

MG

143.3

3star

D

9.4

3star

K

4.4

2star

HB

5.0

4star

M

1.6

2star

T

1.8

4star

MG

175.0

3star

is listed as a midfielder despite playing the majority of 2021 in defense. This hurts his rankings as his defensive play is compared to midfielders who get far more of the ball.

- Turnovers are often attributed to the disposing player when the receiving player clearly stuffs up. This is seen in a lot of the annotated highlights packages on AFL.com.au when a player kicks the ball and the marking player stuffs up the mark. Turnover is marked against the kicker when it isn't their fault.

- Champion Data don't seem to have a system to factor in the general state of the game when decisions are made, such as when and where players are positioned and available to receive the ball. "Effective" disposals can be given even when it's a terrible option based on the state of the game, while "ineffective" disposals can actually lead to goals. Seems wrong

 

[SunshineBoy]

Sam Switkowski is a real low point for this system.

And what's with there being about 12 'elite' general forwards and a handful of 'elite' mids.

Champion Data can try and hide behind whatever ridiculous proprietary rating system they have but this is just plain stupid. Is every year.

Isn’t CD owned by the AFL ? In some way? Would explain sone of the stupidity

 

[juss]

Isn’t CD owned by the AFL ? In some way? Would explain sone of the stupidity

In fairness, every statistical algorithm will throw up some weird results. There's no stat for the "eye test", which is what is needed to provide the qualitative view to the quantitative data.

It'll always be flawed as long as it's purely statistical.

 

[Millky95]

In fairness, every statistical algorithm will throw up some weird results. There's no stat for the "eye test", which is what is needed to provide the qualitative view to the quantitative data.

It'll always be flawed as long as it's purely statistical.

I was working a data conference a few years back and they were talking about how algorithms need humans to go over the data to get anything useful from it. The example they used was a TV station in the USA used an algorithm to pull up interesting stats and the results the algorithm came up with were something like "LA wins most when they score more than the opposition" and "to win, LA need to score more than their opponents". Whilst true, completely unhelpful

 

[owen87]

I was working a data conference a few years back and they were talking about how algorithms need humans to go over the data to get anything useful from it. The example they used was a TV station in the USA used an algorithm to pull up interesting stats and the results the algorithm came up with were something like "LA wins most when they score more than the opposition" and "to win, LA need to score more than their opponents". Whilst true, completely unhelpful

Well I don't know about you but it was certainly helpful new information to me!

 

[WillHayward]

while also batting a gambling addiction, a drug addiction, having anger management issues and being an alcoholic according to the media

Imagine what he can achieve when he gets his s**t together?

Demons could build a footy colony on Mars.

 

[nineteen eighty]

Can't wait for the Dees supporters to jump on and tell us how good Gawn is and that he's elite. If CD say he's not then obviously he's not.

 

[burge13]

Far too many players listed so "elite" isnt the word to use.

Also lol Switkowski > Gawn offical

 

[Johnny Karate]

Can't wait for the Dees supporters to jump on and tell us how good Gawn is and that he's elite. If CD say he's not then obviously he's not.

Imagine how many goals he’ll kick in a Prelim when he’s elite.

 

[Cm Perfect]

Elites v Non-elites, who would win?

"Elites"

B T. Stewart H. Andrews D. Rich
HB J. Crisp D. Moore C. Daniel
C B. Smith C. Petracca H. McCluggage
HF D. Martin L. Franklin D. Zorko
F T. Greene T. Hawkins J. Stringer
R B. Grundy C. Oliver M. Bontempelli
I J. Lever J. Macrae T. Miller D. Parish


Best of the Rest...

B N. Vlastuin S. May D. Grimes
HB C. Salem J. McGovern Luke Ryan
C L. Whitfield O. Wines M. Duncan
HF B. Fritsch A. Naughton T. Papley
F J. Riewoldt H. McKay C. Cameron
R M. Gawn J. Steele R. Laird
I Josh Kelly S. Walsh Z. Merrett T. Taranto


The CD team probably has a superior backline. Best of the Rest has a pretty deep and balanced midfield though. Forward lines about even.

 

[Tap Tap Tapperoo]

You’d think the people at Champion Data would have taken one look at that list and decided to keep the results in house and revisit the algorithm.

 

[dylan93]

Champion Data getting more and more clueless as the years go on.

 

[Strapping Young Lad]

Melbourne fans’ jimmies sure are rustled.

 

[Trav 20]

Melbourne fans’ jimmies sure are rustled.

Considering you have to google what you're meant to think on most topics it's not surprising you can't see the anomalies.

 

[Gysberts2Bate]

Elites v Non-elites, who would win?

"Elites"

B T. Stewart H. Andrews D. Rich
HB J. Crisp D. Moore C. Daniel
C B. Smith C. Petracca H. McCluggage
HF D. Martin L. Franklin D. Zorko
F T. Greene T. Hawkins J. Stringer
R B. Grundy C. Oliver M. Bontempelli
I J. Lever J. Macrae T. Miller D. Parish


Best of the Rest...

B N. Vlastuin S. May D. Grimes
HB C. Salem J. McGovern Luke Ryan
C L. Whitfield O. Wines M. Duncan
HF B. Fritsch A. Naughton T. Papley
F J. Riewoldt H. McKay C. Cameron
R M. Gawn J. Steele R. Laird
I Josh Kelly S. Walsh Z. Merrett T. Taranto


The CD team probably has a superior backline. Best of the Rest has a pretty deep and balanced midfield though. Forward lines about even.

Petracca, Oliver and Bont together would annihilate any midfield they played against, especially if Naitanui came on as ruck to support Grundy. Midfield battle isn’t even close, esp with Gawn well past his best as a tap ruck

 

[groggm]

Elites v Non-elites, who would win?

"Elites"

B T. Stewart H. Andrews D. Rich
HB J. Crisp D. Moore C. Daniel
C B. Smith C. Petracca H. McCluggage
HF D. Martin L. Franklin D. Zorko
F T. Greene T. Hawkins J. Stringer
R B. Grundy C. Oliver M. Bontempelli
I J. Lever J. Macrae T. Miller D. Parish


Best of the Rest...

B N. Vlastuin S. May D. Grimes
HB C. Salem J. McGovern Luke Ryan
C L. Whitfield O. Wines M. Duncan
HF B. Fritsch A. Naughton T. Papley
F J. Riewoldt H. McKay C. Cameron
R M. Gawn J. Steele R. Laird
I Josh Kelly S. Walsh Z. Merrett T. Taranto


The CD team probably has a superior backline. Best of the Rest has a pretty deep and balanced midfield though. Forward lines about even.

Glaring omission in the elite CD team of one S.Switkowski

 

[Dazb86]

The Western with five elite midfielders, I’d hate to come up against them in a pressure situation trying to win centre clearance to get yourself back into the game.

You understand we were only beaten in the middle a couple times last year right. One of those times happened to be the grand final. Doesn't change the fact they outplayed their opponents the vast majority of the year.

 

[WillHayward]

Champion Data need to consult with the clubs over what statistics they value and completely re-do their subjective weighting system.

What the coaches value in a player is one of the most important metrics. So having B&F winners and 5 of the top 10 in the AFLCA award miss the "elite" list is a huge flaw, especially when some of those players are from finals teams.

Aside from that these are some of the clear flaws in their system that I've seen over the years:

- Not adequately rating defensive capabilities, particularly from players who are "lockdown" defenders, not offensive defenders. The ability to stop the opposition and stop goals is clearly valued by the league.

- Not rating uncontested possessions. I can't quite remember where but someone on here found that Champion Data effectively give zero worth to uncontested possessions, when surely having an uncontested possession is better than having no possession at all.

- Giving large penalties for kicking behinds or similar mistakes. For example in this game from Jake Riccardi, Champion Data thought that dropping a mark and having an ineffective handball was worth a greater penalty than the gain from kicking a goal, from a key position forward. That's a flawed weighting system

- Weighting hitouts and hitouts to advantage too highly. Rucks in general are positioned too high in their ratings when there is only a limited correlation between winning these stats and winning games. But then at times dominant around the ground play (eg Max Gawn) is not enough to become "elite"

- Incorrectly assigning positions to players they don't play in. For example at the Crows, Chayce Jones is listed as a midfielder despite playing the majority of 2021 in defense. This hurts his rankings as his defensive play is compared to midfielders who get far more of the ball.

- Turnovers are often attributed to the disposing player when the receiving player clearly stuffs up. This is seen in a lot of the annotated highlights packages on AFL.com.au when a player kicks the ball and the marking player stuffs up the mark. Turnover is marked against the kicker when it isn't their fault.

- Champion Data don't seem to have a system to factor in the general state of the game when decisions are made, such as when and where players are positioned and available to receive the ball. "Effective" disposals can be given even when it's a terrible option based on the state of the game, while "ineffective" disposals can actually lead to goals. Seems wrong

Also stats should be lumped together and viewed holistically. Steele is ranked 2nd in tackles in the AFL and then top 20 in a whole bunch of other things. This makes him obviously "elite" overall.

 

[emblurr]

They've given some insight how they get their high ratings for players

1. How a player wins the ball
2. Where they win the ball
3. What they do with the ball after

It's a reasonable way to rate, as a player that does well in all three categories will impact the game

Can read about it here The best player in the AFL in 2021 “by some margin”, according to Champion Data

 

[Johnny Karate]

Melbourne fans’ jimmies sure are rustled.

Not at all. Imagine how much we’ll improve if we can get some more elite players in…