Mega Thread Nick Daicos - Can he be the GTWEB? Part 2

[Fadge]

Sounds as though poor MR Meow is devastated that his beloved PLaYeR RaTiNGZ algorithm has been exposed for the waste of time that it is.

It's amazing how easy it has been to expose so many fundamental flaws in the algorithm with a little bit of critical thinking.

 

[Dtennent94]

Jagger's 5th ever quarter of AFL was better than any quarter Naicos has ever played

Why hasn't this thread been shut down??

He’s on track to surpass Nick, if he keeps this trajectory up.

 

[Mr Meow]

Sounds as though poor MR Meow is devastated that his beloved PLaYeR RaTiNGZ algorithm has been exposed for the waste of time that it is.

It's amazing how easy it has been to expose so many fundamental flaws in the algorithm with a little bit of critical thinking.

"MR Meow" is adhom nonsense that make you look like a moron.

You have not used critical thinking a single time in this thread, but backslap yourself for each terrible argument you make. Very entertaining.

Player Ratings will be making an even more regular appearance in this thread now that they have been shown to make you cry like a petulant schoolgirl.

 

[threenewpadlocks]

It's just not a tag if they don't follow you. It's just someone playing defensively at stoppages.

It's possible for an opposition coach to tag a player when they're playing midfield but not tag a player when they go forward, going forward because they were so effectively tagged in midfield. Shocking concept, I know, but it's one that does actually happen at AFL level.

 

[Fadge]

Player Ratings will be making an even more regular appearance in this thread now that they have been shown to make you cry like a petulant schoolgirl.

Using casual sexism now? Interesting.

But feel free to continue to use PLaYeR RaTiNGZ in this thread, because anyone who is regularly active in the thread has now been duly informed as to how useless the data is when comparing players.

 

[Fadge]

He’s on track to surpass Nick, if he keeps this trajectory up.

Yet another one to add to the list!

With:
Sheezel
NWM
WAshcroft
Rowell

Etc....

 

[Corn Cobbers]

Using casual sexism now? Interesting.

But feel free to continue to use PLaYeR RaTiNGZ in this thread, because anyone who is regularly active in the thread has now been duly informed as to how useless the data is when comparing players.

Isn't it bedtime now in "London"? Time for someone to pretend to go to bed

 

[Corn Cobbers]

Just a couple more tweaks and PLaYeR RaTiNGZ might actually be useful to someone

 

[Fadge]

Isn't it bedtime now in "London"? Time for someone to pretend to go to bed

I reckon he'll have nightmares about the PLaYeR RaTiNGZ algorithm...

'Press Release from Champion Data:

After much feedback from our stakeholders, the algorithm has been improved to more accurately reflect player performance.

After retrospectively running the algorithm across the past 3 seasons, we are please to advise that Nick Daicos is the clear and undisputed number 1 player in the competition.'

 

[threenewpadlocks]

View attachment 2555441

Just a couple more tweaks and PLaYeR RaTiNGZ might actually be useful to someone

6th best player = game value 0
44th best player = game value 0
Didn't play = game value 0

Confirming that this mathematical formula for COaChEZ VOteZ is correct, given you seem to be the expert on formulas for good or bad games?

 

[Fadge]

6th best player = 0 statistics
44th best player = 0 statistics
Didn't play = 0 statistics.

Confirming that this mathematical formula for COaChEZ VOteZ is correct, given you seem to be the expert on formulas for good or bad games?

Yes, we are all fully aware as to the limitations of coaches votes, and hoping you are now well versed in the (significantly greater) limitations of PLaYeR RaTiNGZ?

I mean, at least the best few players on the ground are always rewarded with coaches votes, as compared to the lottery that is PLaYeR RaTiNGZ.

 

[Corn Cobbers]

Yes, we are all fully aware as to the limitations of coaches votes, and hoping you are now well versed in the (significantly greater) limitations of PLaYeR RaTiNGZ?

I mean, at least the best few players on the ground are always rewarded with coaches votes, as compared to the lottery that is PLaYeR RaTiNGZ.

PLaYeR RaTiNGZ is all they've got at the Whitten Oval, they're almost GOAT contenders there, the way they charge forward manically and try to score as much as possible gives the algorithm a woody

 

[threenewpadlocks]

significantly greater

Why? How? You don't think treating the 6th and 44th best player as having played an identically valuable game is not an even more significant flaw?

I mean, at least the best few players on the ground are always rewarded

We've established that is not true for example the unanimous view by virtually everyone that one of English or Bont played a better game than Bailey Williams in R0.

 

[sr36]

It's possible for an opposition coach to tag a player when they're playing midfield but not tag a player when they go forward, going forward because they were so effectively tagged in midfield. Shocking concept, I know, but it's one that does actually happen at AFL level.

Good luck having your own personal definition, but you must get confused when people including Fagan talk about Brisbane not tagging or briefly trying one, even though Dunkley plays that same defensive role around stoppage every single week.

 

[HeathySure]

Spot on. It's why Castagna was a much better player than Fyfe and Buckley. He was actually a winner.

Castagna was a vital role player for the system, Dustin was the main man. And Lynch and Reiwoldt. Funny how you are making my point for me

 

[HeathySure]

I'd like further elaboration on every point - they were so brilliant I could read dozens of pages dissecting each component further. You were doing so well you almost convinced yourself at one point. Another 5000 words and I think you'll get there.

I dont know why you are replying to everything, Its coming across as smarmy and spiteful. Happens when you cannot refute an argument

 

[doppleganger]

6th best player = game value 0
44th best player = game value 0
Didn't play = game value 0

Confirming that this mathematical formula for COaChEZ VOteZ is correct, given you seem to be the expert on formulas for good or bad games?

Yes the coaches each assess who they think the best 5 players were for each game.

Sure you can be the 10th best player for every game and not get any votes, so it isn't a way of comparing average players.

However the really good players, they are consistently getting coaches votes. Being the best player on the ground will ensure you gets votes.

Score equity ratings tally up a players contribution to scoring, so being part of a high scoring team in a high scoring game will obviously result in larger individual ratings than a player player in a low scoring team.

Score equity ratings effectively reward ALL players playing in a 130 v 115 shootout, over those playing in a 75 v 68 game.

 

[threenewpadlocks]

they are consistently getting coaches votes

No, the good players fail to get coaches votes around one-third of the time.

Being the best player on the ground will ensure you gets votes.

Do you believe being the best player is worth 5 times more than being the 5th best player?

Score equity ratings tally up a players contribution to scoring, so being part of a high scoring team in a high scoring game will obviously result in larger individual ratings than a player player in a low scoring team.

Score equity ratings effectively reward ALL players playing in a 130 v 115 shootout, over those playing in a 75 v 68 game.

You're correct in general, but you're incorrect in the case of the Bulldogs and Collingwood here - in fact in the opposite.

Dogs per game averaged 230 ratings points as a team per game, and the Pies averaged 209. Dogs 10% more.

But it's an equity measurement that is meant track directly to margins of victory, so the fact that the Dogs had a percentage of 137% and Pies had a percntage of 120% (after 2 finals games) is directly relevant here.

But the Dogs were 10% better than the Pies last year purely in points for and points against,, which tracks directly to the fact that they also generated 10% more ratings points as a team.

If the Dogs were a 5% worse each points for and points against team to be a 10% overall worse team, their percentage would have dropped from from 137% to 123% - still above Pies' 120% on the season.

So I just proved you wrong that Dogs generated more points as a team because their games were higher scoring - wrong, they generated more points in direct proportion to their average margins across the season compared to Pies.

Yes, the Dogs did beat up on bad teams more than the Pies, and Bont had a part of it (not that I need reminding), but purely for the mathematics, you're completely, utterly, brazenly incorrect here. Pies fans haven't really made the argument over the last few pages that Bont only appears the more statistical player because he's generating those statistics in junk time minutes and in less meaningful games against bad teams as the Dogs run up the score, but Pies fans haven't made that argument, because it would require admitting that Bont's output is actually superior (even if there's an easier context for generating that output). See, I'm even making Pies' fans arguments for them.

Anyway, so Daicos wasn't harmed at all by the Pies generating 10% fewer ratings points, because Pies were a 10% worse margin team anyway.

So, for Daicos to nonetheless get so few points as Pies' best midfielder, the points have to go somewhere among Pies' games, yes?

It does. To the defenders. The good Collingwood defenders that got the points. Daicos is both not as good as a player because he doesn't undertake defensive acts as often as other midfielders (tackles, pressure acts, one-percenters, intercepts compared to other midfielders who also win their possessions in the defensive half), so the Pies defence is actually highly succeeding despite Daicos per the ratings points.

Indeed, this can be proven by the ranking of player ratings points within all defenders for all Dogs and Pies defenders (min 10 games played):

Dogs rankings: Dale - rank 4 (3)
Lobb - rank 40 (21)
O'Donnell - rank 107 (39)
Bramble - rank 112 (57)
Cleary - rank 115 (75)
Johannisen - rank 121 (93)
Jones - rank 135 (111)
Busslinger - rank 167 (129)

If we had a perfectly 'balanced' defensive group - that is our defenders were equally contributing to our overall 3rd best percentage, you'd expect our best defender to be the 3rd highest ranked defender in the league in general, our 2nd best defender to be the 21st best defender in the league in general (ie, if tracked perfectly to league rankings, each of the 18 teams would have one of the best 18 defenders in order of their percentage on the season, if you follow my logic). Keep tracking those numbers where you +18 for each players, I put in brackets.

Compare to the Pies who had the 5th best percentage, but these were the ratings of all defenders:

J. Daicos - rank 13 (5)
Moore - rank 14 (23)
Maynard - rank 35 (41)
Houston - rank 55 (59)
Howe - rank 60 (77)
Frampton - rank 88 (95)
Perryman - rank 90 (113)
Quaynor - rank 93 (131)

To be absolutely clear, I'm not stating that this is how I would rate the players.

But the points system does give points. And in giving points, Pies had 8 different defenders who averaged more than every single Dogs defender bar two, and Pies had 8 defenders that were better footballers than the Dogs' 3rd best defender on the season (having watched Dogs and Pies games last year, that is absolutely true, irrespective of game style or whatever defence you want to suggest).

Just that, by virtue of the ratings system thought that 8 different Pies defenders were within the top 100 defenders across the whole league in the 2025 season, despite the fact that of 18 teams playing 7-8 defenders a game (plus whoever they drop and promote from reserves) means that there's about 150 "best 22" defenders in the league - its just that Collingwood's depth in defence was so good that player ratings correctly assumes that none of them were in that bottom 50 of the 150 best 22 defenders, which is due to an accurate distribution of Pies' 209 player ratings points per game.

Tl:dr: So I've just proven the Pies fans wrong yet again here. Collingwood got just as many team player ratings points relative to the Dogs as their actual points for and points against percentage was relative to the Dogs. It's not "harder for Pies fans to generate points", it's just that Pies players do generate points - the depth of their defenders, and because Daicos isn't the player largely contributing to the Pies very good defensive record, he doesn't get the ratings points, and therefore isn't actually that good of a player.

 

[sr36]

Tl:dr: So I've just proven the Pies fans wrong yet again here. Collingwood got just as many team player ratings points relative to the Dogs as their actual points for and points against percentage was relative to the Dogs. It's not "harder for Pies fans to generate points", it's just that Pies players do generate points - the depth of their defenders, and because Daicos isn't the player largely contributing to the Pies very good defensive record, he doesn't get the ratings points, and therefore isn't actually that good of a player.

No you've proven yourself wrong. You've shown what we've all been saying. Bont gets high ratings because he's a gun player in a high scoring team. Meanwhile Collingwood players, despite being a better team than the Dogs last year, didn't score as high on player ratings, because they were a lower scoring team.

 

[Fadge]

Do you believe being the best player is worth 5 times more than being the 5th best player?

It's certainly better than the best player on the ground receiving fewer ranking points than 20 other players on the ground, and in many cases, half as many ranking points as another player who wouldn't have been in the best 10 on the ground in the same game, or the best 15 on the ground in a higher scoring game.

Take Daicos in Round 0.

10 coaches votes.
Second best player across the entire round as per Garry Lyon
Clearly the best and most influential player on the ground according to all sane judges.

Player Rating of 9.3.

Jack Sinclair was rated as nearly 3 x better. De Goey nearly twice as good. TDK more than 50% better.

And here are the players from the same round who were apparently more than twice as good as Daicos:

 

[threenewpadlocks]

It's certainly better than the best player on the ground receiving fewer ranking points than 20 other players on the ground,

Good thing we're taking a sample over 20+ games to average out the formula. You looooooooove to point out that Daicos got 25+ ratings points in the Prelim loss to the Lions without getting any coaches votes. You can't have it both ways if this is your argument. But you've previously pointed out in this thread, numerous times, that Daicos did actually play well in finals last year depsite his 0 coaches votes (and supported by the ratings points).

You really can't have it both ways wit your argument - you're entirley contradicting yourself.

 

[Fadge]

Good thing we're taking a sample over 20+ games to average out the formula.

As I've said earlier, it's not basic maths where two negatives make a positive.

If individual games are wrong, the total data set is wrong.

 

[Fadge]

You looooooooove to point out that Daicos got 25+ ratings points in the Prelim loss to the Lions without getting any coaches votes. You can't have it both ways if this is your argument. But you've previously pointed out in this thread, numerous times, that Daicos did actually play well in finals last year depsite his 0 coaches votes (and supported by the ratings points).

You really can't have it both ways wit your argument - you're entirley contradicting yourself.

Nah, I take the mickey with my reference to Daicos in the Preliminary Final.

One of his highest rated games of his career, and wouldn't be in his best 50% of games.

Because the algorithm is broken.

 

[threenewpadlocks]

No you've proven yourself wrong. You've shown what we've all been saying. Bont gets high ratings because he's a gun player in a high scoring team. Meanwhile Collingwood players, despite being a better team than the Dogs last year, didn't score as high on player ratings, because they were a lower scoring team.

Did you actually read out my post that actively proved this statement wrong - with actual mathematics - yet you're posting it after the fact?

Huh? Utterly incredible. I literally just proved this statement wrong and you're ... literally repeating it verbatim, again.

In summary, because you seem incapable of understanding very, very basic concepts

Ratings points = track margins of victory

Dogs 137%, Pies 120%. Dogs 10% better team (purely in average margin). Dogs percentage would be very close to 120% as well if they were a 5% worse team in each of points for and points against (and therefore a 10% worse overall team).

Dogs ratings points 230 per game, Pies ratings points 209 per game (10% more)

ie, they scale identically

ie, the points are distributed within the team for how good they were

ie, how high or low scoring dogs/pies to each other is completely cancelled out and plays no factor to how the Pies and Dogs as a team collectively generated ratings points.

ie, the ratings points for the fact that the Pies were a good team and generated 209 ratings points per game on average went to the players that were actually good and contributed to victory - the fact that the Pies have 8 top 100 defenders in the league, the Dogs had only 2 top 100 defenders in the league. Daicos contributed to Pies success far less than what you'd think a top 1 player in the league would suggest, because maybe he isn't actually as good as the top 1 player in the league.

 

[Fadge]

ie, the points are distributed within the team for how good they were

Nek Minit....

Pies finished 4th and played in a Preliminary Final, whilst the Dogs finished 9th, and missed finals.

Tell us again how much better the Doggies were than the Pies?

10% you reckon?