Opinion Who is currently the best player in the AFL

[JHF1870]

Obviously. I'm not saying supercoach points, ratings points or anything are always a good approximation for player-game-value. Indeed, there are times that I think I could do better, upon immediately having attended a game live and watched them.

Just that at least they're there, and accessible, and better than anything I could do using stats that are publicly available, if I wanted to give a value to the 130,000+ player-games since 2012.

All good, wasn't disagreeing with anything you said, just adding to it.

 

[threenewpadlocks]

Has been explained numerous times, I rate the coaches view point of who the best players were each game over player ratings.

I disagree with this view because

• coaches do not rate how well a player played, even when they're expected to be the best 5 players, when they do not play as the best 5 on the ground - leaving 41 behind
• teams play better, and win by bigger margins, but points values are fixed. GWS players got all 20 votes for a 27 point win over hawthorn. Dogs players got all 20 votes for a 86 point win over GWS, including playing better games than all the players who had just gotten 20 votes than the week before. Do you think that the Dogs' team collectively played better in that win over GWS by a bigger margin than GWS played collectively over Hawthorn the week before?

I don't necessarily disagree that it is useful information to know which 5 players in order the coaches thought played the best game. In fact, I agree with you that it is a very useful thing to know, and in many cases, could be more accurate than if we to rank the same 5 players according to the rankings points they generated within that game.

I just don't think it's very useful to determine the overall value or goodness of players as measured by their accumulation over all games they play, given the above two dot points.

We don't know which of Lachie Blakiston or Peter Wright the coaches thought was the better player for Essendon vs. us. Nobody has asked them of that.

Do YOU know where Luke Beveridge or Chris Fagan ranked Marcus Bontempelli within the 46 players of the game in Round 0? I rate that view too, but we simply don't have it.

 

[RCAB]

How do you weigh a CC vs an SC on defensive half?

Or do you assume SC evenly balanced between front and back half.

I'm adding or subtracting on this years data at least; 7 points or so but that data suggests 2-3 point variance.

I don't.. I just pull the data from wheeloratings. I guess only CD would have those stats.

You may be right that SC stats might look different in different parts of the ground.

 

[RCAB]

Has been explained numerous times, I rate the coaches view point of who the best players were each game over player ratings.

The coaches use player ratings and CD to form opinions and ... assign coaches votes.

 

[Fadge]

The coaches use player ratings and CD to form opinions and ... assign coaches votes.

Like the normal, informed person who understands football, 95% of their assessment will be based on what they saw, and 5% based on STaTZ.

 

[JHF1870]

I don't.. I just pull the data from wheeloratings. I guess only CD would have those stats.

You may be right that SC stats might look different in different parts of the ground.

I think i can improve the stats by basing on the team stats for efficiency

 

[RCAB]

RCAB I'm only doing this part time to keep myself busy and that I'm a bit of a nerd. I do know the limitations in general and in my system. I'm looking into improving the system at least for next season and providing more context.

It's going to take 3-4 times the time investment to do however. It's already quite onerous.

Last night Adelaide had 59 inside 50s for 17 goals. Carlton had 60 for 12. I think given the team value and impact and at least being able to remove the crude negative loss bonus I would reduce I50 values for Carlton players over Adelaide or at least have some weighted system to say inside 50s were poor return for Carlton so Carltons I50 scores would be lower than Adelaides.

So what you want is:

Adjusted I50 value = player inside 50s × team scoring efficiency from inside 50s

Right?

Similar thing below

Same for Clearances, Clearance value is indicitave forward pressure on the scoreboard. High clearances and Low scoreboard pressure means the values should be lower.

At least this way stats will be based more on team output.

Adjusted clearance value = player clearances × team scoreboard return from clearances?

Even saying this I understand the limitations that Walsh could have had 100% impact on the scoreboard from clearances but the rest of the team could have been crap so I'm basically just dealing with team averages unless I were to go even further and I really don't have that sort of time.

I understand what you're saying.

You adjust by using team averages as a proxy for individual value.

So if Carlton waste entries, a player who actually delivered great inside 50s could still get unfairly dragged down because the team as a whole converted badly.

That makes your model more contextually aware but could also be more
team contaminated and therefore noisy.

My model is different as it correlates by proximity the player stats that historically line up better with premiership winning football.

It's more like:

1. Use the player stats that matter
2. Standardise them
3. Weight them
4. Sum them into one score

For each weighted stat, I converted the player’s raw value into an AFL-wide percentile within that season.

So if a player was better than most of the pool in MetresGained, they got a high percentile.

In other words, ModelScore = the sum of each weighted stat percentile.

So the model is not based on raw totals alone. Each stat is first converted into an AFL-wide percentile so different types of output can be compared on the same scale, and then the weights are applied.

 

[doppleganger]

The coaches use player ratings and CD to form opinions and ... assign coaches votes.

Yes Roby, the coaches looked at the player ratings after the 2025 PF and saw Daicos had a player rating of 26 which was almost 10 higher than the 2nd best player on the ground, and then chose to ignore the player ratings when giving their votes.

 

[Mr Meow]

Yes Roby, the coaches looked at the player ratings after the 2025 PF and saw Daicos had a player rating of 26 which was almost 10 higher than the 2nd best player on the ground, and then chose to ignore the player ratings when giving their votes.

See, you are obsessed with referencing Player Ratings. That's the whole reason I trust them. If someone references the same positive Daicos rating from 1 game on 1000 different occasions, then it must be a valuable metric.

 

[threenewpadlocks]

Yes Roby, the coaches looked at the player ratings after the 2025 PF and saw Daicos had a player rating of 26 which was almost 10 higher than the 2nd best player on the ground, and then chose to ignore the player ratings when giving their votes.

But they absolutely didn't look at the player ratings for the Round 3, 2026 win over GWS, where Daicos was rank 1, before deciding that Daicos was BOG? How can you know either way, if your evidence is merely that player ratings are sometimes disjointed to coaches votes - but what about the times they align, like three rounds ago?

 

[RCAB]

Yes Roby, the coaches looked at the player ratings after the 2025 PF and saw Daicos had a player rating of 26 which was almost 10 higher than the 2nd best player on the ground, and then chose to ignore the player ratings when giving their votes.

You don't even know who buys the product. You don't know anything. You just post on an internet forum.

 

[Fadge]

See, you are obsessed with referencing Player Ratings. That's the whole reason I trust them. If someone references the same positive Daicos rating from 1 game on 1000 different occasions, then it must be a valuable metric.

That game, amongst many others, has been referenced to support the fact the Player Ratings algorithm simply doesn't work.

But you do you.

 

[Fadge]

But they absolutely didn't look at the player ratings for the Round 3, 2026 win over GWS, where Daicos was rank 1, before deciding that Daicos was BOG? How can you know either way, if your evidence is merely that player ratings are sometimes disjointed to coaches votes - but what about the times they align, like three rounds ago?

Desperate reach.

Woohoo for Player Ratings!

There is the rare occasion they are accurate!

 

[Fadge]

You don't even know who buys the product. You don't know anything. You just post on an internet forum.

We do know the algorithm is simply mo better than a novice darts player throwing darts at a darts board.

 

[Mr Meow]

That game, amongst many others, has been referenced to support the fact the Player Ratings algorithm simply doesn't work.

But you do you.

Is there a reason you only cherry pick examples where coaches votes and Player Ratings aren't in alignment?

Is there a reason you don't do that with coaches vs Brownlow votes, coaches vs League MVP ratings, and specific BnF vs coaches votes discrepancies on a game by game basis?

Using coaches votes/Brownlow votes suggests that the 41 or 43 players the umpire/coach doesn't rate, all had the same quality of game. That is, if you use them to rank players seasons on game by game output. That doesn't really "work" either.

 

[Mr Meow]

We do know the algorithm is simply mo better than a novice darts player throwing darts at a darts board.

Proof?

 

[Mr Meow]

Desperate reach.

Woohoo for Player Ratings!

There is the rare occasion they are accurate!

How are you defining rare, as opposed to sometimes or mostly? Gut feel?

 

[Fadge]

Here are some examples of games this year:

Where are the rest of the votes? The 22nd best 'rated' player, Errol Gulden, received 8 coaches votes in this game.


What? No votes for the 2nd, 3rd and 4th highest 'rated' player on the ground?


Lots of votes missing here. The coaches were unanimous in voting the 19th highest rated player, N. Daicos, as the best on ground.


Unanimous coaches BOG again outside the top 10 highest rated players on the ground. Equal second best player (Cripps) the 25th highest rated.


Just throw a dart in this game....



I had to look a long way down the list for the coaches unanimous BOG in this game! I found Harley Reid ranked 34th, with a Player Rating of 5.6...


Throw a dart in this game. Equal BOG Will Ashcroft the 19th rated player.

View attachment 2578831
Complete and utter raffle.

I think everyone gets the picture...

 

[Fadge]

Using coaches votes/Brownlow votes suggests that the 41 or 43 players the umpire/coach doesn't rate, all had the same quality of game. That is, if you use them to rank players seasons on game by game output. That doesn't really "work" either.

Never go full Padlock....

 

[Mr Meow]

Never go full Padlock....

So you have no valid reply to that.

7 of the top 8 ranked players by Player Ratings on Easter Monday received coaches votes. Is there a reason you didn't comment on that?

Rachele was easily the highest rated on Thursday, then Berry and Walsh. Why aren't you referencing that?

There is confimation bias to a level that is normal for most people. Then there is the extreme level you take it to, on Planet Fadge.

 

[Mr Meow]

Here are some examples of games this year:
View attachment 2578823
Where are the rest of the votes? The 22nd best 'rated' player, Errol Gulden, received 8 coaches votes in this game.

View attachment 2578824
What? No votes for the 2nd, 3rd and 4th highest 'rated' player on the ground?

View attachment 2578825
Lots of votes missing here. The coaches were unanimous in voting the 19th highest rated player, N. Daicos, as the best on ground.

View attachment 2578826
Unanimous coaches BOG again outside the top 10 highest rated players on the ground. Equal second best player (Cripps) the 25th highest rated.

View attachment 2578827
Just throw a dart in this game....


View attachment 2578829
I had to look a long way down the list for the coaches unanimous BOG in this game! I found Harley Reid ranked 34th, with a Player Rating of 5.6...

View attachment 2578830
Throw a dart in this game. Equal BOG Will Ashcroft the 19th rated player.

View attachment 2578831
Complete and utter raffle.

I think everyone gets the picture...

Is there a reason you've excluded dozens of other games, and within the ones you've chosen, only highlighted there is not alignment?

How is this meant to verify your claim that they "rarely" get it right?

 

[Fadge]

So you have no valid reply to that.

As per the previous 20 replies to such a desperate reach, Brownlow and coaches votes don't assert to 'rate' every player on the ground.

They are a means to find the best player across the season in the eyes of the coaches and umpires, and subsequently award the Brownlow Medal or Coaches Award accordingly.

7 of the top 8 ranked players by Player Ratings on Easter Monday received coaches votes. Is there a reason you didn't comment on that?

Rachele was easily the highest rated on Thursday, then Berry and Walsh. Why aren't you referencing that?

There is confimation bias to a level that is normal for most people. Then there is the extreme level you take it to, on Planet Fadge.

I did say there are occasions that the coaches votes are aligned to highest rated players, in the same way there are occasions that I hit triple 20 when playing darts.

 

[Mr Meow]

Here are some examples of games this year:
View attachment 2578823
Where are the rest of the votes? The 22nd best 'rated' player, Errol Gulden, received 8 coaches votes in this game.

View attachment 2578824
What? No votes for the 2nd, 3rd and 4th highest 'rated' player on the ground?

View attachment 2578825
Lots of votes missing here. The coaches were unanimous in voting the 19th highest rated player, N. Daicos, as the best on ground.

View attachment 2578826
Unanimous coaches BOG again outside the top 10 highest rated players on the ground. Equal second best player (Cripps) the 25th highest rated.

View attachment 2578827
Just throw a dart in this game....


View attachment 2578829
I had to look a long way down the list for the coaches unanimous BOG in this game! I found Harley Reid ranked 34th, with a Player Rating of 5.6...

View attachment 2578830
Throw a dart in this game. Equal BOG Will Ashcroft the 19th rated player.

View attachment 2578831
Complete and utter raffle.

I think everyone gets the picture...

Round 4 Player Ratings scores for highest coaches votes players:

8+ votes

24.3
17.6
13.6
18.2
16.7
20.2
20.7
23.5
24.4
11.3
15.0

4-7 votes

15.3
15.7
22.0
4.7
14.6
15.0
3.7
15.0
9.9
10.1
17.6
11.1
15.0
15.4
16.5
18.5
8.2

A few outliers, but most seemed to score very well.

30 of the top 50 ranked players by Player Ratings received coaches votes.

55 of 368 players received coaches votes for the round: about 15%.

Your dartboard or "rarely" theory suggests that Player Ratings might pick out 8 or less coaches votes recipients from it's top 50 players of the round. Yet here we have 30.

Let me know if any of tjhe basic maths and logic goes over your head.

 

[Fadge]

Round 4 Player Ratings scores for highest coaches votes players:

8+ votes

24.3
17.6
13.6
18.2
16.7
20.2
20.7
23.5
24.4
11.3
15.0

4-7 votes

15.3
15.7
22.0
4.7
14.6
15.0
3.7
15.0
9.9
10.1
17.6
11.1
15.0
15.4
16.5
18.5
8.2

A few outliers, but most seemed to score very well.

32 of the top 50 ranked players by Player Ratings received coaches votes.

55 of 368 players received coaches votes for the round: about 15%.

Your dartboard or "rarely" theory suggests that Player Ratings might pick out 8 or less coaches votes recipients from it's top 50 players of the round. Yet here we have 32.

Let me know if any of tjhe basic maths and logic goes over your head.

My favourite is Harley Reid being ranked the 34th highest 'rated' player, despite being a unanimous BOG by the coaches...

 

[Mr Meow]

My favourite is Harley Reid being ranked the 34th highest 'rated' player, despite being a unanimous BOG by the coaches...

So again, your whole theory of them "rarely" getting it right is by cherry picking whichever examples suit your agenda, then ignoring the rest.

I showed a strong correlation between coaches votes and the higher ranked players.