GeorgeFloydOverdosed said: Jeremy Doku plays alongside some of the best at Man City, and yet is supposedly a dud 48.12% in spite of 18 acc, 17 pace.
I think even if you spent a while looking at this guy's stats, you'd have to conclude there's something wrong with that low 48.12% rating..
..and yet Doku gets a measly 6.9 average rating before being sold off. Expand
So what is it in this case that is dragging Doku down? I see he has the 10 temperament and good natural fitness, pace, acceleration. Is it mostly the consistency?
LightningFlik said: Does this lad get 90 minutes or come off the bench? Expand Mostly full appearances
Rain said: So what is it in this case that is dragging Doku down? I see he has the 10 temperament and good natural fitness, pace, acceleration. Is it mostly the consistency? Expand
We can see that Doku is lacking in anticipation, concentration, jumping reach, and is a bit lower on a few physicals. Rashford also has significantly better hiddens & personality, but Doku's personality/hiddens aren't particularly bad.
@GeorgeFloydOverdosed I've been thinking recently about an experiment to measure the importance of attributes for a given role (e.g. Striker); could you tell me if you think this is worthwhile?
Create a test league (maybe 10 or so teams) with identical coaching staffs (Assistant Manager is probably sufficient, since he'll be the one managing games), and identical players in all positions. That is, all the teams' goalkeepers are identical (copied from a real player), as are their fullbacks, their centrebacks and so on. None of the players ought to be able to play the position you're testing (e.g. Striker).
All teams are set up to play the same formation (if this can't be done in the pre-game editor, it would require ten human managers on holiday, hence identical Assistant Managers). All players have their attributes and conditions frozen.
The experiment is taking real-world strikers (e.g. Haaland, Mbappe, Vini, Goncalo Ramos) and putting multiple copies of them on to different teams, meaning all teams are identical except for their different strikers. If they're all playing the same tactics with the same coaches, the only variation will come from their strikers.
And I don't even think you necessarily need to look at the league table at the end of the experiment to determine which striker is better, since games can be won and lost by other players too. I'd look at the goals scored by the strikers, their assist count and possibly their average match ratings (or other things like key passes, xG, if you don't trust how match ratings are generated).
From here, you would perform a linear regression to assess what attributes / personality values are important in affecting the things we care about. I have to think that if was useful, someone would have done it by now but I can't see what's wrong with it.
LightningFlik said: @GeorgeFloydOverdosed I've been thinking recently about an experiment to measure the importance of attributes for a given role (e.g. Striker); could you tell me if you think this is worthwhile?
Create a test league (maybe 10 or so teams) with identical coaching staffs (Assistant Manager is probably sufficient, since he'll be the one managing games), and identical players in all positions. That is, all the teams' goalkeepers are identical (copied from a real player), as are their fullbacks, their centrebacks and so on. None of the players ought to be able to play the position you're testing (e.g. Striker).
All teams are set up to play the same formation (if this can't be done in the pre-game editor, it would require ten human managers on holiday, hence identical Assistant Managers). All players have their attributes and conditions frozen.
The experiment is taking real-world strikers (e.g. Haaland, Mbappe, Vini, Goncalo Ramos) and putting multiple copies of them on to different teams, meaning all teams are identical except for their different strikers. If they're all playing the same tactics with the same coaches, the only variation will come from their strikers.
And I don't even think you necessarily need to look at the league table at the end of the experiment to determine which striker is better, since games can be won and lost by other players too. I'd look at the goals scored by the strikers, their assist count and possibly their average match ratings (or other things like key passes, xG, if you don't trust how match ratings are generated).
From here, you would perform a linear regression to assess what attributes / personality values are important in affecting the things we care about. I have to think that if was useful, someone would have done it by now but I can't see what's wrong with it. Expand My view is that doing tests in situ is better than these artificial simulations. For example, if you make every player have flat 10 in everything, then the results would show you that Mbappe does worse than Ramos because both have pace/acc far exceeding 10, but Ramos has better mentals than Mbappe. I don't know if it would actually happen that way, but it's just an example to illustrate what can go wrong. Also if you freeze morale, condition, etc. then that impairs the effect of temperament, natural fitness, and so on. I can think of 100 ways things could wrong, and that's just what we know.. what if there's things we don't know about, such as rock-paper-scissors situations that occur in standard gameplay?
It's like taking this guy and sticking him in the Ukraine war:
Yes, he's the world's best finisher under controlled conditions, but context matters. And the following isn't part of my response to you, but I just want to point out that we don't need to send this guy to Ukraine 1000+ times to have an inkling - clone and send him in 5-10 times, and we'll end up with a pretty solid idea.
If you just test for what actually works in a real league, with standard play conditions, then you can skip all that and yet know what works for certain.
What I'm currently trying to grapple with is how to weight pace/acc vs. the rest of what matters. Neither alone works. Pace/acc is more all-or-nothing, so you can't weight it at '15' like concentration '13' is weighted. Filtering pace/acc at ~14 instead of weighting it works, but then you end up with Ramos being better than Mbappe which isn't really what we want. And it doesn't help that ratings aren't a reliably solid measure of actual performance. And we also have to consider that pace/acc gains through training of +4 are achievable.
As touched on before, we also have to consider that someone like Ramos probably does great because his league (I only assume here) generally has low pace/acc to allow him to thrive. If you stuck him in the Premier league, the situation would switch where high pace/acc is required and therefore dominates over high mentals/technicals. So dealing with apples & oranges discrepancies as well as possible.
As I'm testing to find the right balance, I realize that Genie Scout seems to have something f'd up going on with how it calculates ratings again. Might try FMRTE for now to see if the calculation is different.
GeorgeFloydOverdosed said: Jeremy Doku plays alongside some of the best at Man City, and yet is supposedly a dud 48.12% in spite of 18 acc, 17 pace.
I think even if you spent a while looking at this guy's stats, you'd have to conclude there's something wrong with that low 48.12% rating..
..and yet Doku gets a measly 6.9 average rating before being sold off.
So what is it in this case that is dragging Doku down? I see he has the 10 temperament and good natural fitness, pace, acceleration. Is it mostly the consistency?
LightningFlik said: Does this lad get 90 minutes or come off the bench?
Mostly full appearances
Rain said: So what is it in this case that is dragging Doku down? I see he has the 10 temperament and good natural fitness, pace, acceleration. Is it mostly the consistency?
We can see that Doku is lacking in anticipation, concentration, jumping reach, and is a bit lower on a few physicals. Rashford also has significantly better hiddens & personality, but Doku's personality/hiddens aren't particularly bad.
@GeorgeFloydOverdosed I've been thinking recently about an experiment to measure the importance of attributes for a given role (e.g. Striker); could you tell me if you think this is worthwhile?
Create a test league (maybe 10 or so teams) with identical coaching staffs (Assistant Manager is probably sufficient, since he'll be the one managing games), and identical players in all positions. That is, all the teams' goalkeepers are identical (copied from a real player), as are their fullbacks, their centrebacks and so on. None of the players ought to be able to play the position you're testing (e.g. Striker).
All teams are set up to play the same formation (if this can't be done in the pre-game editor, it would require ten human managers on holiday, hence identical Assistant Managers). All players have their attributes and conditions frozen.
The experiment is taking real-world strikers (e.g. Haaland, Mbappe, Vini, Goncalo Ramos) and putting multiple copies of them on to different teams, meaning all teams are identical except for their different strikers. If they're all playing the same tactics with the same coaches, the only variation will come from their strikers.
And I don't even think you necessarily need to look at the league table at the end of the experiment to determine which striker is better, since games can be won and lost by other players too. I'd look at the goals scored by the strikers, their assist count and possibly their average match ratings (or other things like key passes, xG, if you don't trust how match ratings are generated).
From here, you would perform a linear regression to assess what attributes / personality values are important in affecting the things we care about. I have to think that if was useful, someone would have done it by now but I can't see what's wrong with it.
LightningFlik said: @GeorgeFloydOverdosed I've been thinking recently about an experiment to measure the importance of attributes for a given role (e.g. Striker); could you tell me if you think this is worthwhile?
Create a test league (maybe 10 or so teams) with identical coaching staffs (Assistant Manager is probably sufficient, since he'll be the one managing games), and identical players in all positions. That is, all the teams' goalkeepers are identical (copied from a real player), as are their fullbacks, their centrebacks and so on. None of the players ought to be able to play the position you're testing (e.g. Striker).
All teams are set up to play the same formation (if this can't be done in the pre-game editor, it would require ten human managers on holiday, hence identical Assistant Managers). All players have their attributes and conditions frozen.
The experiment is taking real-world strikers (e.g. Haaland, Mbappe, Vini, Goncalo Ramos) and putting multiple copies of them on to different teams, meaning all teams are identical except for their different strikers. If they're all playing the same tactics with the same coaches, the only variation will come from their strikers.
And I don't even think you necessarily need to look at the league table at the end of the experiment to determine which striker is better, since games can be won and lost by other players too. I'd look at the goals scored by the strikers, their assist count and possibly their average match ratings (or other things like key passes, xG, if you don't trust how match ratings are generated).
From here, you would perform a linear regression to assess what attributes / personality values are important in affecting the things we care about. I have to think that if was useful, someone would have done it by now but I can't see what's wrong with it.
My view is that doing tests in situ is better than these artificial simulations. For example, if you make every player have flat 10 in everything, then the results would show you that Mbappe does worse than Ramos because both have pace/acc far exceeding 10, but Ramos has better mentals than Mbappe. I don't know if it would actually happen that way, but it's just an example to illustrate what can go wrong. Also if you freeze morale, condition, etc. then that impairs the effect of temperament, natural fitness, and so on. I can think of 100 ways things could wrong, and that's just what we know.. what if there's things we don't know about, such as rock-paper-scissors situations that occur in standard gameplay?
It's like taking this guy and sticking him in the Ukraine war:
Yes, he's the world's best finisher under controlled conditions, but context matters. And the following isn't part of my response to you, but I just want to point out that we don't need to send this guy to Ukraine 1000+ times to have an inkling - clone and send him in 5-10 times, and we'll end up with a pretty solid idea.
If you just test for what actually works in a real league, with standard play conditions, then you can skip all that and yet know what works for certain.
What I'm currently trying to grapple with is how to weight pace/acc vs. the rest of what matters. Neither alone works. Pace/acc is more all-or-nothing, so you can't weight it at '15' like concentration '13' is weighted. Filtering pace/acc at ~14 instead of weighting it works, but then you end up with Ramos being better than Mbappe which isn't really what we want. And it doesn't help that ratings aren't a reliably solid measure of actual performance. And we also have to consider that pace/acc gains through training of +4 are achievable.
As touched on before, we also have to consider that someone like Ramos probably does great because his league (I only assume here) generally has low pace/acc to allow him to thrive. If you stuck him in the Premier league, the situation would switch where high pace/acc is required and therefore dominates over high mentals/technicals. So dealing with apples & oranges discrepancies as well as possible.
As I'm testing to find the right balance, I realize that Genie Scout seems to have something f'd up going on with how it calculates ratings again. Might try FMRTE for now to see if the calculation is different.