LightningFlik said: I've always suspected that the coach report is vibes-based, really. It makes sense to me that their current reputation (and possibly even form) influences it, as well as the relative strength of the league and team-mates. I don't for one minute think the game is telling me that a 4.5 star player is unquestionably better than a 4 star player. Expand Now I get the weird star ratings my meta team gets. Even though they have equal attributes, the star ratings are all over the place because of different CA. Setting them to 1 CA, they're all equally 3.5 stars.
For normal play, I think this means that wingers are going to be overrated by your staff (i.e. 4 star because of high CA cost for attributes), whereas DM will be underrated (i.e. 2 star because of low CA cost of attributes). Both of these players are equal quality in reality.
I came across one of my notes in regards to player reputation:
Home reputation influences national team call-ups by coaches, but not the selection of starters/substitutes at match time
I'm saving these notes for one big compendium sometime, but since it was raised, thought some of you might be interested in knowing that one.
I've done some more tests of Teamwork 8 > 20: 6th (sacked), 13th (sacked), 4th, 7th (sacked) = 7.5 position
So giving all outfield players teamwork is not good. But giving the wingers teamwork is seemingly good. I suspect this could be due to what HarvestGreen identified as certain attributes being less beneficial the more players that use them, rather than it specifically being a winger thing that Mustermann finds. But who knows.
I realized that I'm being too hasty in doing away with sprints & pressures. I got swept up in what Mustermann had to say, that I completely forgot about what the game files actually say:
| Distance run | Work Rate, Stamina, Pace | | Sprints | Work Rate, Stamina | | Pressures attempted | Work Rate, Anticipation | | Pressure completion | Work Rate, Anticipation, Acceleration |
Idk, having no control and extremely low sample size on this kind of test is kinda invalid statistically. Could just be the effect of rng itself. But just for curiosity then its probably fine.
I took a look at correlations between the top players using my weightings and the various stat measurements, and the one thing that seemed to be consistent was the inconsistency. So while two did stand out as fairly robust correlates, 'interceptions' and 'fouls committed', overall I suspect that these various stats actually have nothing to do with the win or loss outcome. My guess is that its calculating based on certain attributes and tactics or whatnot, independent of the actual match engine outcome. So cosmetic basically. Just a theory.
I tested HarvestGreen's claim that certain attributes have benefits if one player has them instead of the whole team. I'm sure I must have tested this before, since it's important if true, but I still find that it doesn't improve things. Actually, giving several players each one of the several attributes he listed as '20', seemed make the results worse than standard: 10th, 11th, 5th, 12th. This could be because attributes such as decisions are known to have a negative effect, or it could be another reason. So, unknown cause, but its not inexplicable.
The thing is, HarvestGreen's data has always turned to be generally correct for me, so long as you take it in the confines of what he is measuring. So this claim of his must be true under certain conditions. And the teamwork result (2 wingers having +8 teamwork improves team result, but whole team at +12 teamwork doesn't) seems to be evidence this kind of thing exists. Obviously it probably has something to do with tactics and/or attribute distributions. Whatever it is, it doesn't seem like you can reduce it to 'chuck in this one guy with 20 decisions'.
I decided to take a look at HarvestGreen's excel data on it, not expecting anything interesting but thought I should probably take a quick look to see if anything further could be gleaned. I saw that he included a table I don't think he mentioned in his posts, which I've translated:
So it looks like he's actually tested some of those key stat measures. It's pretty apparent to me that what these results are showing is that the attributes involved matter, not the combo of them as 'dribbles' or 'interceptions' or whatnot.
LightningFlik said: I've always suspected that the coach report is vibes-based, really. It makes sense to me that their current reputation (and possibly even form) influences it, as well as the relative strength of the league and team-mates. I don't for one minute think the game is telling me that a 4.5 star player is unquestionably better than a 4 star player.
Now I get the weird star ratings my meta team gets. Even though they have equal attributes, the star ratings are all over the place because of different CA. Setting them to 1 CA, they're all equally 3.5 stars.
For normal play, I think this means that wingers are going to be overrated by your staff (i.e. 4 star because of high CA cost for attributes), whereas DM will be underrated (i.e. 2 star because of low CA cost of attributes). Both of these players are equal quality in reality.
I came across one of my notes in regards to player reputation:
Home reputation influences national team call-ups by coaches, but not the selection of starters/substitutes at match time
I'm saving these notes for one big compendium sometime, but since it was raised, thought some of you might be interested in knowing that one.
I've done some more tests of Teamwork 8 > 20: 6th (sacked), 13th (sacked), 4th, 7th (sacked) = 7.5 position
So giving all outfield players teamwork is not good. But giving the wingers teamwork is seemingly good. I suspect this could be due to what HarvestGreen identified as certain attributes being less beneficial the more players that use them, rather than it specifically being a winger thing that Mustermann finds. But who knows.
I realized that I'm being too hasty in doing away with sprints & pressures. I got swept up in what Mustermann had to say, that I completely forgot about what the game files actually say:
| Distance run | Work Rate, Stamina, Pace |
| Sprints | Work Rate, Stamina |
| Pressures attempted | Work Rate, Anticipation |
| Pressure completion | Work Rate, Anticipation, Acceleration |
I've got some ideas I'm going to look into now.
Possession tactic with less technical players
Idk, having no control and extremely low sample size on this kind of test is kinda invalid statistically. Could just be the effect of rng itself. But just for curiosity then its probably fine.
I took a look at correlations between the top players using my weightings and the various stat measurements, and the one thing that seemed to be consistent was the inconsistency. So while two did stand out as fairly robust correlates, 'interceptions' and 'fouls committed', overall I suspect that these various stats actually have nothing to do with the win or loss outcome. My guess is that its calculating based on certain attributes and tactics or whatnot, independent of the actual match engine outcome. So cosmetic basically. Just a theory.
I tested HarvestGreen's claim that certain attributes have benefits if one player has them instead of the whole team. I'm sure I must have tested this before, since it's important if true, but I still find that it doesn't improve things. Actually, giving several players each one of the several attributes he listed as '20', seemed make the results worse than standard: 10th, 11th, 5th, 12th. This could be because attributes such as decisions are known to have a negative effect, or it could be another reason. So, unknown cause, but its not inexplicable.
The thing is, HarvestGreen's data has always turned to be generally correct for me, so long as you take it in the confines of what he is measuring. So this claim of his must be true under certain conditions. And the teamwork result (2 wingers having +8 teamwork improves team result, but whole team at +12 teamwork doesn't) seems to be evidence this kind of thing exists. Obviously it probably has something to do with tactics and/or attribute distributions. Whatever it is, it doesn't seem like you can reduce it to 'chuck in this one guy with 20 decisions'.
I decided to take a look at HarvestGreen's excel data on it, not expecting anything interesting but thought I should probably take a quick look to see if anything further could be gleaned. I saw that he included a table I don't think he mentioned in his posts, which I've translated:
So it looks like he's actually tested some of those key stat measures. It's pretty apparent to me that what these results are showing is that the attributes involved matter, not the combo of them as 'dribbles' or 'interceptions' or whatnot.
Premier League 2.1 weights
@GeorgeFloydOverdosed
Crazy CB 88,6 Rating