Opinion Who is currently the best player in the AFL

Fadge · Apr 10, 2026

RCAB said:
I've also used some standardized modelling that correlates metrics that better reflect winning football and premiership-level impact. For example... some stats constantly show up as high for teams that win premierships or reach the GF (you can go check Wheeloratings) each year. So using that logic, you try to see which players correlate to those metrics.

Yes, Champion Data via wheeloratings - 'premiership-level impact' - where 9 of their top 11 players in 2025 failed to win a final, and the other 2 won a single final.

Can't argue with 'premiership-level impact' as formulated by the CD algorithm!

JHF1870 · Apr 10, 2026

RCAB said:
Aren't you reacting to small sample size data if you adjust per start of the year in round 4 or earlier? How do you know it's not merely a short trend that will revert to the mean later in the season?

Well this is last years result (18 game minimum)

I experimented with changes in Finals and tweaked again when Thilthorpe broke the system a little which sometimes can be unavoidable. Defenders can now get really good scores but they don't do it as often.

I made some significant changes to balance positions out. Center clearances are worth more than stoppage and goals more points. I've also added a winning/losing bias which was more of a Brownlow predictor thing but looking at it so far it hasn't made a huge difference only for players like Zac Merrett and Tim Taranto.

Well that's the thing with LDU, is he going to maintain his score against stronger opponents? Will Treacy be able to maintain his big scores?

Bont was still my #1 player last year comfortably.

DaicosDisciple35 · Apr 10, 2026

The correct answer is also the most boring one. It changes day to day because they are all humans and not robots.

I think the top 2 are pretty clear in Daicos and Bont - they are the most consistent, skillful and impactful players in the comp on a week to week basis. From all reports something that separates them both from the herd is their off-field preparation and dedication to being elite athletes.

Outside of those I would list the following as just as likely to BOG on any given day as anyone

in no order:
Jordan Dawson
Kosiah Pickett
Max Gawn
Isaac Heenyt
Errol Gulden
Finn Callaghan
Will Ashcroft

I think slightly below that tier you have guys like
Chad Warner
Nick Watson
Bailey Smith
Max Holmes
Jeremy Cameron
Jordan De Goey
Hugh Mcluggage
Josh Dunkley
Zac Butters
Sam Darcy
Noah Andersen
Matt Rowell
Christian Petraaca
Toby Greene
Brody Grundy
Ed Richards

There are plenty of players I left off this second list but again it changes a lot.

JHF1870 · Apr 10, 2026

RCAB said:
Aren't you reacting to small sample size data if you adjust per start of the year in round 4 or earlier? How do you know it's not merely a short trend that will revert to the mean later in the season?

Oh, if you meant the 1/3 of inside 50s, it was a general estimate. That's why went lower but also because sometimes inside 50s can be farmed junk stats.

DaicosDisciple35 · Apr 10, 2026

JHF1870 said:
Well this is last years result (18 game minimum)

I experimented with changes in Finals and tweaked again when Thilthorpe broke the system a little which sometimes can be unavoidable. Defenders can now get really good scores but they don't do it as often.

I made some significant changes to balance positions out. Center clearances are worth more than stoppage and goals more points. I've also added a winning/losing bias which was more of a Brownlow predictor thing but looking at it so far it hasn't made a huge difference only for players like Zac Merrett and Tim Taranto.

Well that's the thing with LDU, is he going to maintain his score against stronger opponents? Will Treacy be able to maintain his big scores?

Bont was still my #1 player last year comfortably.

View attachment 2578129

This is a pretty good list, not sure anything can ever really measure just how much better a kick bont, daicos and gulden are compared to the rest of the comp. They do things other players can't even dream of and it changes the entire dynamic of the game.

RCAB · Apr 10, 2026

Fadge said:
Yes, Champion Data via wheeloratings - 'premiership-level impact' - where 9 of their top 11 players in 2025 failed to win a final, and the other 2 won a single final.

Can't argue with 'premiership-level impact' as formulated by the CD algorithm!

It's a genius move to align yourself with this line of posting doppleganger

JHF1870 · Apr 10, 2026

DaicosDisciple35 said:
This is a pretty good list, not sure anything can ever really measure just how much better a kick bont, daicos and gulden are compared to the rest of the comp. They do things other players can't even dream of and it changes the entire dynamic of the game.

Thanks.

I wanted to build the system initially to predict the Brownlow but seems impossible even with losing bias given the way it's handled. I also wanted to build a system that could remove any perceived bias by viewing and I've argued before that the more a certain player gets media hype the more likely we are to vote them the best player. If we're busy watching Gulden, Heeney and Warner, we miss McInerney.

Also to add to previous post I measure impact and have revised a lot since last year in which I understood that Center clearances are one of the main chains that lead to goal and while I can't tell whether a stoppage clearance lead to an inside 50, i'll assume it's worth 1/2 that of a centre clearance.

The inside 50s last night was 30% for Crows and 20% for Carlton and I'm using a very conservative score for inside 50s relative to goals.

The thing I've always argued is that the #1 priority is to score a goal. You can evade 3 people and pinpoint a pass to a player in the corridor but the end result is the same with a contested disposal in any other situation. Yes it looks fantastic and shows the brilliance of the player but realistically how do we measure that? Footy isn't always sexy and just like the NBA the flashiest players don't always win championships.

The game has some brilliant kicks. I think Finn Callaghan is elite top 5 kicks in the competition.

threenewpadlocks · Apr 10, 2026

doppleganger said:
Just like all your ramblings about player ratings, I have no interest in how many ratings points you attribute to players who are yet to play a game.

Bont plays. We don't pick Luke Kennedy
Bont is late out. We do pick Luke Kennedy.
Bont's value when he does play is therefore how much better of a footballer he is than Luke Kennedy.
But we have to attribute some sort of rating to Luke Kennedy, because that's the baseline that we're measuring Bont against to determine the value of Bont to the Bulldogs via his selection availability.
If you have no interest in how well Luke Kennedy is likely to play should he be selected this week, you, by extension, have no interest in how valuable Marcus Bontempelli is as a footballer.
Luke Kennedy will continue to be a player yet to play a game after this weekend, but the idea of Luke Kennedy playing, because he would have replaced Bont, is necessary to understand the value of Bont's fitness.
Therefore, we have to attribute something, some sort of expected rating, of Kennedy's game, had he played.

Any issue with the above?

Fadge · Apr 10, 2026

JHF1870 said:
Well this is last years result (18 game minimum)

I experimented with changes in Finals and tweaked again when Thilthorpe broke the system a little which sometimes can be unavoidable. Defenders can now get really good scores but they don't do it as often.

I made some significant changes to balance positions out. Center clearances are worth more than stoppage and goals more points. I've also added a winning/losing bias which was more of a Brownlow predictor thing but looking at it so far it hasn't made a huge difference only for players like Zac Merrett and Tim Taranto.

Well that's the thing with LDU, is he going to maintain his score against stronger opponents? Will Treacy be able to maintain his big scores?

Bont was still my #1 player last year comfortably.

View attachment 2578129

So '18 game minimum' is the threshold?

So an 18 game player is compared without any deductions to a player who played 25, 26, or 27 games, but a 17 game player is not eligible?

Interesting data analysis decision.

RCAB · Apr 10, 2026

JHF1870 said:
Well this is last years result (18 game minimum)

I experimented with changes in Finals and tweaked again when Thilthorpe broke the system a little which sometimes can be unavoidable. Defenders can now get really good scores but they don't do it as often.

I made some significant changes to balance positions out. Center clearances are worth more than stoppage and goals more points. I've also added a winning/losing bias which was more of a Brownlow predictor thing but looking at it so far it hasn't made a huge difference only for players like Zac Merrett and Tim Taranto.

Well that's the thing with LDU, is he going to maintain his score against stronger opponents? Will Treacy be able to maintain his big scores?

Bont was still my #1 player last year comfortably.

View attachment 2578129

Can you explain a bit more when you mean TT broke the system?

So in terms of correlation from 2022-25, about 8% relationship strength vs
stoppage clearances 15% (based on teams closer to a premiership n those years).

So centre clearances while meaningful they're not a very good way of predicting who plays the best premiership-style football during that period.

The correlation has gone up in 2024-25 from 2023-22 (but it did drop down from 2024 to 2025 even though we had the same premiers). And stoppage clearances always appear as the stronger correlation over CC.

doppleganger · Apr 10, 2026

threenewpadlocks said:
Bont plays. We don't pick Luke Kennedy

Bont is late out. We do pick Luke Kennedy.

Bont's value when he does play is therefore how much better of a footballer he is than Luke Kennedy.

But we have to attribute some sort of rating to Luke Kennedy, because that's the baseline that we're measuring Bont against to determine the value of Bont to the Bulldogs via his selection availability.

If you have no interest in how well Luke Kennedy is likely to play should he be selected this week, you, by extension, have no interest in how valuable Marcus Bontempelli is as a footballer.

Luke Kennedy will continue to be a player yet to play a game after this weekend, but the idea of Luke Kennedy playing, because he would have replaced Bont, is necessary to understand the value of Bont's fitness.

Therefore, we have to attribute something, some sort of expected rating, of Kennedy's game, had he played.

Any issue with the above?

It is not an individual game like baseball where you can easily judge a players individual performance through pitching and batting stats.

AFL is a team game with 23 players that rotate through different roles and positions.

When Daicos was a late out against the Lions, the emergency E.Allan didn't simply take the role as Pies #1 attacking midfielder who would be at all centre bounces.

It meant that other senior players like Sidey, J.Daicos, Crisp and DeGoey are asked to spend more time as midfielders.

DeGoey, Crisp and Sidey were all awful - game player ratings of 2 or below - is that also attributable to Daicos not playing? Does Daicos help let those guys play more complementary roles that better suits there game?

So many issues with you simplistic one out one in approach, which is compounded by the scoreboard contribution ratings themselves which are basically saying playing in a team that wins 70-62 has poorly rated players compared to a team that loses 104-120.

threenewpadlocks · Apr 10, 2026

doppleganger said:
It is not an individual game like baseball where you can easily judge a players individual performance through pitching and batting stats

I don't dispute the difficulty of using statistics to measure player value and goodness when compared to Baseball. That changing roles and players within a team absorbing each other's role makes that a challenging task.

But beyond that, you can agree with me that Bontempelli offers positive value to the Bulldogs in that we are avaliable to select him over Luke Kennedy?

You can agree that that's also true for Daicos and Allan for this weekend too?

And at least attempting to approximate this and represent it by some sort of numerical value is at least a worthwhile exercise? So that we can compare Bont, Daicos, Kennedy and Allan?

Otherwise what's the point? We're circling back to "but my eyes tell me Daicos>Bont" argument which is all just a bit stupid.

At least I'm trying. I'm giving you my reasoning. You're just rejecting it out of hand and pointing out the errors in my reasoning without offering up your own attempt to approximate this and have some sort of comparative value between the 4 players.

I don't think it's particularly good faith to continually criticise my methodology (if not the principle), without going to the effort of attempting an approximation, a numerical value yourself.

As the saying goes, it's easy to criticise, it's hard to actually hard to create and do yourself.

JHF1870 · Apr 10, 2026

RCAB said:
Can you explain a bit more when you mean TT broke the system?

So in terms of correlation from 2022-25, about 8% relationship strength vs
stoppage clearances 15% (based on teams closer to a premiership n those years).

So centre clearances while meaningful they're not a very good way of predicting who plays the best premiership-style football during that period.

The correlation has gone up in 2024-25 from 2023-22 (but it did drop down from 2024 to 2025 even though we had the same premiers). And stoppage clearances always appear as the stronger correlation over CC.

That was my finals experiment testing things out and the stats for him were overvalued.

threenewpadlocks · Apr 10, 2026

doppleganger said:
DeGoey, Crisp and Sidey were all awful - game player ratings of 2 or below - is that also attributable to Daicos not playing?

Perhaps, but it's a small sample (of one game, for those three players), and a player's performance in one game can naturally vary around their true talent level.

Which is why I wouldn't use Gallagher's actual average performance in the three games he played in 2025, I would use an estimation of his true talent level, that is, a true talent level that can be approximated by scoring around 5.5 ratings points per game.

And in any case, I'm not necessarily interested in the value gained from Bontempelli playing instead of Gallagher or Kennedy. I'm interested in how good Bontempelli is as a player in general, that is, the value he would offer to a league average team, therefore how much better he is than a league average replacement level player. Which is itself not a real thing, just conceptual in nature.

JHF1870 · Apr 10, 2026

RCAB said:
Can you explain a bit more when you mean TT broke the system?

So in terms of correlation from 2022-25, about 8% relationship strength vs
stoppage clearances 15% (based on teams closer to a premiership n those years).

So centre clearances while meaningful they're not a very good way of predicting who plays the best premiership-style football during that period.

The correlation has gone up in 2024-25 from 2023-22 (but it did drop down from 2024 to 2025 even though we had the same premiers). And stoppage clearances always appear as the stronger correlation over CC.

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.

doppleganger · Apr 10, 2026

threenewpadlocks said:
Otherwise what's the point? We're circling back to "but my eyes tell me Daicos>Bont" argument which is all just a bit stupid.

At least I'm trying. I'm giving you my reasoning.

I don't know why you keep trying, that is the point.

It took you way to long to even accept that you didn't even understand the workings of the player ratings before conceding that you had it wrong.

And now you are just blabbering about "value" and again I disagree with your approach and desire to appoint a numerical value to it.

Pies ins this week (Daicos and Pendles) are +12.5, but you think Bont is more than that on his own!?

Bont is a star

, he plays for a high scoring fun team to watch in the Dogs, and is thus a "score contribution ratings" darling.

threenewpadlocks said:
As the saying goes, it's easy to criticise, it's hard to actually hard to create and do yourself.

Are you claiming you created the player ratings?

hr_23 · Apr 10, 2026

Kosi Pickett best player in the game atm

threenewpadlocks · Apr 10, 2026

doppleganger said:
I don't know why you keep trying, that is the point.

Because this thread title is "who is currently the best player in the AFL".

That is, how valuable they are on an average AFL team.

That is, how much better they are than an average AFL replacement level player.

I am merely trying to address that question via numerical approximation, to allow for comparison between players.

If you think that Ratings Points are such a poor thing to use for the basis of a numerical approximation, that's your belief, that I would disagree with, and the reasons I disagree with I think are valid (as it is better at predicting players' likely contributions to future outcomes than supercoach points).

doppleganger said:
Pies ins this week (Daicos and Pendles) are +12.5, but you think Bont is more than that on his own!?

Not at all.

According to wheelo, the expected ratings points for Pendles and Daicos is 11.2 and 15.8. Lets say, because of the Pies slower play style (that you always claim), that a Pies replacement player would get fewer points than a Dogs player, who gets more points in an attacking team - 4.0 instead of 5.5. Value = 19.0 (7.2+11.8)

Bont expected 19.3, above Dogs replacement player 5.5, value is 13.8

19.0 > 13.8

Daicos and Pendles together worth more than Bont.

doppleganger said:
Are you claiming you created the player ratings?

I'm claiming I created the idea that 5.5 is a replacement level, or that 4.0 is Collingwood's replacement level, or that even a week or two ago that I decided to adjust the ratings points (with your prompting) for high and low scoring games the manner I did.

And again, given you haven't disagreed with my reasoning to at least attempt to numerically compare Daicos and Bont using this logic, do you have a better method? What different numbers or approximation of game-value would you use other than Player Ratings?

JHF1870 · Apr 10, 2026

hr_23 said:
Kosi Pickett best player in the game atm

One of the most exciting that s for sure. Hard to hate him.

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.

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.

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.

doppleganger · Apr 10, 2026

threenewpadlocks said:
Because this thread title is "who is currently the best player in the AFL".

Yep.

threenewpadlocks said:
That is, how valuable they are on an average AFL team.

That is, how much better they are than an average AFL replacement level player.

Not sure many go down this path of thinking when are who is the best player.

threenewpadlocks said:
If you think that Ratings Points are such a poor thing to use for the basis of a numerical approximation, that's your belief, that I would disagree with, and the reasons I disagree with I think are valid (as it is better at predicting players' likely contributions to future outcomes than supercoach points).

Oh yes they were 0.015 points better, but getting worse, according to an individual tipping model assessment.

threenewpadlocks said:
Lets say, because of the Pies slower play style (that you always claim), that a Pies replacement player would get fewer points than a Dogs playerwho gets more points in an attacking team

Yeah but that isn't the approach taken.
Wheelo just gives Anderson (our debutant) a rating of 5.7.

threenewpadlocks said:
I'm claiming I created the idea that 5.5 is a replacement level, or that 4.0 is Collingwood's replacement level, or that even a week or two ago that I decided to adjust the ratings points (with your prompting) for high and low scoring games the manner I did.

So you just take other people's work, interpret some of it poorly and then create some poor charts.

Well done.

threenewpadlocks said:
And again, given you haven't disagreed with my reasoning to at least attempt to numerically compare Daicos and Bont using this logic, do you have a better method? What different numbers or approximation of game-value would you use other than Player Ratings?

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

threenewpadlocks · Apr 10, 2026

We can run the same logic using supercoach score, if we so wish. Supercoach adjusts for pace of play. 3300 points are awarded to each AFL game.

Personally, I think supercoach points are worse across the entire league for all players than ratings points to value how good a player can be expected to be in the future, but we can use them all the same.

Bont = averaging 141 points
Daicos = averaging 131 points
replacement level = about 45 points, eyeballing it, maybe better ways of measuring this
Bont = 96 points above replacement
Daicos = 86 points above replacement

10 more supercoach points = 1 more scoreboard point

ie, Bont worth 9.6 points per game for playing
Daicos worth 8.6 points per game for playing

Incidentally (well no, because the same logic is used by bookmakers), this is approximately how much bookmaker lines move when there is a late out, assuming that the possibility of a late out wasn't already priced into the market. Top 5 player in the league out about 9 points, best player in team/AA out: 6-7 points, elite player 4-5 points, established good best 23 top 15 player 2-3 points, good fringe player 1-2 points. etc.

threenewpadlocks · Apr 10, 2026

doppleganger said:
Yeah but that isn't the approach taken.
Wheelo just gives Anderson (our debutant) a rating of 5.7.

Dude, I can disagree with Wheelo on the basis for his prediction and use 5.5 (or 4.0) instead, and have valid (if disagreeable) logic for doing so.

You'll see he predicts 5.7 for Louis Emmett too, and all debutants. Irrespective of the team they play for, or the position they play in.

If anything, adjusting for all teams is an improvement you make.

JHF1870 · Apr 10, 2026

threenewpadlocks said:
We can run the same logic using supercoach score, if we so wish. Supercoach adjusts for pace of play. 3300 points are awarded to each AFL game.

Personally, I think supercoach points are worse across the entire league for all players than ratings points to value how good a player can be expected to be in the future, but we can use them all the same.

Bont = averaging 141 points
Daicos = averaging 131 points
replacement level = about 45 points, eyeballing it, maybe better ways of measuring this
Bont = 96 points above replacement
Daicos = 86 points above replacement

10 more supercoach points = 1 more scoreboard point

ie, Bont worth 9.6 points per game for playing
Daicos worth 8.6 points per game for playing

Incidentally (well no, because the same logic is used by bookmakers), this is approximately how much bookmaker lines move when there is a late out, assuming that the possibility of a late out wasn't already priced into the market. Top 5 player in the league out about 9 points, best player in team/AA out: 6-7 points, elite player 4-5 points, established good best 23 top 15 player 2-3 points, good fringe player 1-2 points. etc.

Problem with any stat based system is certain players like D'Ambrosio can farm points very easily.

I've experienced this with my system. Somehow Butts scored very high against you guys earlier this season but his game was sub standard over Freijeh and Bont.

threenewpadlocks · Apr 10, 2026

doppleganger said:
So you just take other people's work, interpret some of it poorly and then create some poor charts.

I don't think "just taking other people's work" (your characterisiation) for what I'm trying to do is fair no more than it would be than the fact that you know, in general research and science, people are obviously always building upon the work of others provided they ethically reference it.

Karl Jackson's PhD (then of Champion Data),that invented ratings points, didn't invent the idea of equity in the AFL, that was Darren O'Shaughnessy, the head numbers guy at the Saints now and then of Champion Data too when he did the field equity stuff (Jackson took his job effectively, but under different CD ownership).

If you're going to characterise my attempts to use the work of others to do more work myself as "taking" it, you're just being a dick to me, and you're showing your true colours here.

threenewpadlocks · Apr 10, 2026

JHF1870 said:
Problem with any stat based system is certain players like D'Ambrosio can farm points very easily.

I've experienced this with my system. Somehow Butts scored very high against you guys earlier this season but his game was sub standard over Freijeh and Bont.

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.

Log In or Create New Account

Rd 9 Reviews + Club Forums »

Opinion Who is currently the best player in the AFL

Who is currently the best player in the AFL

Marcus Bontempelli

Nik Daicos

Sam Darcy

Nasiah Wanganeen-Milera

Errol Gulden

Jeremy Cameron

Zak Butters

Isaac Heeney

Matt Rowell

Other

Christian Petracca

Kysaiah pickett

Hall of Famer

Premium Platinum

Premium Platinum

Premium Platinum

Premium Platinum

Norm Smith Medallist

Premium Platinum

Brownlow Medallist

Hall of Famer

Norm Smith Medallist

Brownlow Medallist

Brownlow Medallist

Premium Platinum

Brownlow Medallist

Premium Platinum

Brownlow Medallist

Club Legend

Brownlow Medallist

Premium Platinum

Brownlow Medallist

Brownlow Medallist

Brownlow Medallist

Premium Platinum

Brownlow Medallist

Brownlow Medallist

Similar threads