Yarema said: I feel there is still some hidden information. Specifically I feel that some effects are immediate as in the day before intake, some before preview and some (and this might be a bit tinfoil hat theory) might take 2-3 seasons to show. Expand I've done some testing of this, just to rule out the possibility that reloading my saves just prior to intake wasn't affecting results of PA testing I was doing. There was no difference, for reloading just before intake as compared to starting new save from scratch each time.
EvensenFM said: Thanks for this! I think I read somewhere else recently (probably Reddit) that Youth Facilities don't impact the PA of newly generated players at all. This is really good to know. Expand
I actually posted a comment on your youtube recently, that was it.
You do pretty good videos, enjoying watching them alongside doing things, even though I wouldn't have thought football history would be my thing.
kvasir said: Great test, man. Thoughts about the affiliate clubs thing:
There are 7 countries with a higher Youth Rating than England(136), highest being Brazil with 163. If my top English club gets an affiliate in Brazil, wouldn't that be a good thing? Seems like swapping an English newgen for a Brazilian should be an upgrade most of the time.
The question is, I guess, how the game makes the Brazilian kid: does he get generated in my youth intake with my own 20/20 JC and YR? or is he made at his own club, with their stats (probably lower than mine), and just moved to my team?
If it's the first option then its a win. If not, then its a gamble. Any ideas? Expand
Yes, you are entirely correct in your first hypothesis. I just removed the affiliates from Man City because I'm trying to work out the newgen mechanics accurately.
To the question, a few days ago I would have said to you that its just Brazil's youth rating + it must be your own facilities, because in some testing I did before on it I found that it only takes the youth rating, not any of the affiliate club attributes.
But recently I was watching EBFM's video on affiliate clubs, and his data shows that while nation youth rating is still the most important factor, affiliate club attributes do in fact influence the quality of the newgen. So now I'm not sure, but I'd say believe what EBFM says on this because it's likely I only took like 5 samples or something in my test, because intake from affiliate clubs isn't really a part of the formula for newgen mechanics that I've been testing for.
tl;dr 'Youth Facilities' have no effect on newgen PA in any way. The 'club reputation' of OTHER clubs in your nation has a small effect on newgen PA. 'Corporate facilities' may also have a minor effect on newgen PA, but this could be just statistical noise.
----------------------------
I did testing before in FM19 and found that youth facilities affect newgen CA but not PA.
I have done a decent amount of testing with FM24 and have found the same, and that median PA continues to be the most reliable measure of newgen quality, with peak PA having much more of a randomness factor.
What I found in my previous testing was roughly the following:
Junior coaching ~40% PA effect Nation youth rating ~25% PA effect Youth recruitment ~25% PA effect Club reputation acts as tie breaker for youth recruitment Either unique nation or division ID ~25% (or perhaps even more) PA effect (a hidden, unchangeable factor - it may be 'Nation attribute template' that is listed as an inaccessible debug option in the editor, which sounds like it would be similar to the 'Nation personality template' before it was removed in recent versions) All other factors no effect
But I wanted to test a few new theories I had this time as well.
First, the data:
Man City normal (no affiliates), all other England teams normal: 141.5 median, 170 peak Man City normal (no affiliates), all other England teams YF1: 138.1 median, 161.4 peak Man City normal (no affiliates), all other England teams YF20: 143.7 median, 171.4 peak Man City normal (no affiliates), all other England teams JC1: 143 median, 163 peak Man City normal (no affiliates), all other England teams ClubRep/Train/YouthImp/Corp 1: 124.3 median, 176.3 peak Man City normal (no affiliates), all other England teams Train/YouthImp/Corp 1: 137.25 median, 165.5 peak Man City normal (no affiliates), all other England teams YouthImp/Corp 1: 136 median, 171.6 peak Man City normal (no affiliates), all other England teams YouthImp 20/Corp 1: 135.6 median, 164.9 peak Man City normal (no affiliates), all other England teams ClubRep1: 130.6 median, 177 peak Man City normal (no affiliates), all other England teams ClubRep1/YF1: 142 median, 174 peak Man City normal (no affiliates), premier league normal, all other England teams ClubRep1/YF1/corp1: 138 median, 163.5 peak Man City normal (no affiliates), premier league ClubRep/YF1/Corp1, all other clubs normal: 138.8 median, 177.4 peak Man City normal (no affiliates), north-west local region normal, all other clubs ClubRep1/YF1: 138.4 median, 176 peak Man City normal (no affiliates), all other England teams ClubRep/Corp 1: 133.9 median, 167.3 peak
Margin of error I'd guesstimate at ~3-5 median PA. This isn't up to EBFM's standards, but it's 'good enough' in my opinion. I've always been of the view that if you can't replicate it reliably in a few seasons of FM, then its not significant enough of a result to include anyway.
In case you're wondering, keeping the affiliates for Man City reduces median PA by ~15-20, and I deliberately removed them because I know affiliates mess with the newgen intake results as they can replace one or more of them with one from an affiliate club. I wanted to get a pure English club result.
I noticed that Man City often gets some purely Scotland/Ireland only players, and that once their top player (194 PA) was Ireland only. This is notable because Ireland has low enough of a nation youth rating that a 194 PA Ireland player is pretty highly unlikely. It gave credence to the idea that newgens are tied to local region in some way, so I thought I'd re-examine that aspect too (I'd previously tested 'local region' and found it had zero effect on CA/PA).
Hypothesis: Youth facilities have insignificant effect at a single club, but a nations or divisions teams YF contributes cumulatively to a pool of PA that clubs then draw upon. Conclusion: False.
Hypothesis: Youth facilities don't effect median PA, but they effect peak PA or the skew. Conclusion: False.
Although YF1 always eventually produced a high peak PA, I did notice that YF1 alone had a long run of ~155-160 peak PA. This wasn't the case in other YF1 tests, and YF20 produced peak PAs all over the place from 159 to 194. Skew is difficult to definitively conclude on, but I didn't get the impression looking through the results that youth facilities effected the skew.
Hypothesis: Youth facilities contribute to a division pool only. Conclusion: False.
Hypothesis: Youth facilities contribute to a 'local region' pool only. Conclusion: False.
Hypothesis: Junior coaching/Training Facilities/Youth Importance of other clubs has an effect somehow. Conclusion: False.
Hypothesis: Club reputation of other clubs has an effect somehow. Conclusion: Tentatively, yes.
This was a somewhat surprising finding. From memory, I did find before that club rep had a ~10% CA effect, but not on PA. In this case, we can see that ClubRep1 reduced median PA by ~7.7% alone, and ~12.2% when combined with other low club attributes. I took extra samples to be sure and got the same result.
The strange thing is that ClubRep1 result is normal when combined with YF1. I think there's a clue as to what's going on in the fact that the low median PA results only happened when *all* other England teams had low club rep. But that still doesn't make sense of it.
Hypothesis: Corporate facilities of other clubs has an effect somehow. Conclusion: Unclear.
Corp 1 did produce lower median PAs consistently, but the difference of ~3-7 may not be statistically significant. Consider that the typical variance of a club each year is ~5 median PA, even after stripping away the outliers.
Hypothesis: Starting CA affects PA, say that +20 CA due to factor contributes +20 PA as well. Conclusion: False.
SaMaHaJoGu said: What was the difference in his different Long Throws? Expand Long throws must have increased from 1 lot of 'set piece routines' in my 'mixed' screenshot. Even 'match practice' doesn't seem to do it, as long throws is not a highlighted role attribute. Sessions do seem to train the attributes they say they do. My theory was that sessions with low priority for outfield players, like set pieces, have greater overall growth (which turned out to be wrong).
SaMaHaJoGu said: If you do further testing for set pieces, please do help fill the rest of us in. Expand I'm making another post here soon that will say my conclusion about set piece attributes (and surmising everything else), which is that they're pretty much useless and it doesn't matter even if your best takers have only say '6' attribute.
For example:
A season may have ~10 penalties. The difference between '6' and '16' penalty taking is perhaps ~30% success rate, so 3 goals total. Top teams probably give away less penalties due to better 'decisions', so realistically this may be just 1 key goal over the season - which may be 1 key game win. However this would no doubt be outweighed by the CA cost or rareness of the high 'penalty taking' player. Only ~30% of good striker options age 20 or under have 10 or higher penalty taking. Therefore it should be considered a luxury and a tie-breaker between 2 equal ~180+ PA players.
HarvestGreen22's data shows us that +6 pace = +40 goals. We can guess +10 penalty taking = +3 goals. So +10 penalty taking = +0.45 pace. So you're better off glancing at the pace/acc/dribbling stats and choosing the one that has 1 extra pace/acc/dribbling.
The Zippo data also suggestively shows corners and free kicks only giving +1 season goal each from +12 each.
Two other ways it can be expressed is 1 consistency = 1.25 set piece = 3 important matches, but this is just a fun piece of trivia not useful for deciding between players.
If you do want a good set piece taker for comfort, best to buy one instead of training up one.
I observed in the spreadsheets that set pieces decline -1.00 even when you have CA growth up to 15 with the blue 'general' training sessions. By contrast there is 0 decline when you use many of the other session types.
So I came up with some theories.
The first was that perhaps instead of set pieces & technicals in general being squeezed out by low CA, it was instead the 'general' sessions being usually 80-100% priority while the other sessions are 20-60% priority - perhaps the the remaining 40-80% was directed to a hidden pool of attribute growth, just like how 'rest' increases physicals by default. This hidden attribute growth would be significantly more overall growth, which makes sense, as in games focusing on specific attributes often carries an efficiency cost as a game balance.
Perhaps 'priority' can be stacked up to 100%, but anything more simply changes the proportions of that 100%, and each extra session is an extra session of hidden attribute growth. So for instance, 'attacking' (80%) + 'overall' (100%) = 2 x hidden growth and 100% directed growth, whereas 'goalkeeping' (20%) + 'penalties' (20%) + 'outfield' (60%) = 3 x hidden growth and 100% directed growth. This would make sense from what we know SI does, where they use misdirection to keep the true mechanics unknown - in this case, the best training regime would be what appears the least likely and simplistic.
But I did some rough tests, and can rule all of the above out. HarvestGreen22's Quick + Attack + Rest + Quick focus produced significantly better results with less injuries.
I thought I'd share some screenshots, as this shows what you can expect in reality as opposed to a perfect environment averaged 1000 times - one season, England Div 3, 3-4 star coaches, a generous amount of good but not amazing youngsters actually training at the club, just let the chips fall where they may in terms match experience, injuries and so forth (but I'm still selecting amongst the 'survivors' to show here).
Walters_quickattack = Quick + Attack + Rest + Quick focus. 40 injuries. Walters_mixed = My dense but optimized regime based on theories above. 80 injuries. Walters_extreme = Complete overload on physicals training + 2 x match practice & 'overall' to boost technicals/mentals. 207 injuries.
So in typical play is the HarvestGreen22 training significantly better? Yes, it gives 1-2 extra Pace/Acc in 1 season, halves the injury rate, and might even have an edge in mentals over a more 'balanced' regime
I've done some testing of this, just to rule out the possibility that reloading my saves just prior to intake wasn't affecting results of PA testing I was doing. There was no difference, for reloading just before intake as compared to starting new save from scratch each time.
I actually posted a comment on your youtube recently, that was it.
You do pretty good videos, enjoying watching them alongside doing things, even though I wouldn't have thought football history would be my thing.
kvasir said: Great test, man. Thoughts about the affiliate clubs thing:
There are 7 countries with a higher Youth Rating than England(136), highest being Brazil with 163. If my top English club gets an affiliate in Brazil, wouldn't that be a good thing? Seems like swapping an English newgen for a Brazilian should be an upgrade most of the time.
The question is, I guess, how the game makes the Brazilian kid: does he get generated in my youth intake with my own 20/20 JC and YR? or is he made at his own club, with their stats (probably lower than mine), and just moved to my team?
If it's the first option then its a win. If not, then its a gamble. Any ideas?
Yes, you are entirely correct in your first hypothesis. I just removed the affiliates from Man City because I'm trying to work out the newgen mechanics accurately.
To the question, a few days ago I would have said to you that its just Brazil's youth rating + it must be your own facilities, because in some testing I did before on it I found that it only takes the youth rating, not any of the affiliate club attributes.
But recently I was watching EBFM's video on affiliate clubs, and his data shows that while nation youth rating is still the most important factor, affiliate club attributes do in fact influence the quality of the newgen. So now I'm not sure, but I'd say believe what EBFM says on this because it's likely I only took like 5 samples or something in my test, because intake from affiliate clubs isn't really a part of the formula for newgen mechanics that I've been testing for.
----------------------------
I did testing before in FM19 and found that youth facilities affect newgen CA but not PA.
I have done a decent amount of testing with FM24 and have found the same, and that median PA continues to be the most reliable measure of newgen quality, with peak PA having much more of a randomness factor.
What I found in my previous testing was roughly the following:
Junior coaching ~40% PA effect
Nation youth rating ~25% PA effect
Youth recruitment ~25% PA effect
Club reputation acts as tie breaker for youth recruitment
Either unique nation or division ID ~25% (or perhaps even more) PA effect (a hidden, unchangeable factor - it may be 'Nation attribute template' that is listed as an inaccessible debug option in the editor, which sounds like it would be similar to the 'Nation personality template' before it was removed in recent versions)
All other factors no effect
But I wanted to test a few new theories I had this time as well.
First, the data:
Man City normal (no affiliates), all other England teams normal: 141.5 median, 170 peak
Man City normal (no affiliates), all other England teams YF1: 138.1 median, 161.4 peak
Man City normal (no affiliates), all other England teams YF20: 143.7 median, 171.4 peak
Man City normal (no affiliates), all other England teams JC1: 143 median, 163 peak
Man City normal (no affiliates), all other England teams ClubRep/Train/YouthImp/Corp 1: 124.3 median, 176.3 peak
Man City normal (no affiliates), all other England teams Train/YouthImp/Corp 1: 137.25 median, 165.5 peak
Man City normal (no affiliates), all other England teams YouthImp/Corp 1: 136 median, 171.6 peak
Man City normal (no affiliates), all other England teams YouthImp 20/Corp 1: 135.6 median, 164.9 peak
Man City normal (no affiliates), all other England teams ClubRep1: 130.6 median, 177 peak
Man City normal (no affiliates), all other England teams ClubRep1/YF1: 142 median, 174 peak
Man City normal (no affiliates), premier league normal, all other England teams ClubRep1/YF1/corp1: 138 median, 163.5 peak
Man City normal (no affiliates), premier league ClubRep/YF1/Corp1, all other clubs normal: 138.8 median, 177.4 peak
Man City normal (no affiliates), north-west local region normal, all other clubs ClubRep1/YF1: 138.4 median, 176 peak
Man City normal (no affiliates), all other England teams ClubRep/Corp 1: 133.9 median, 167.3 peak
Margin of error I'd guesstimate at ~3-5 median PA. This isn't up to EBFM's standards, but it's 'good enough' in my opinion. I've always been of the view that if you can't replicate it reliably in a few seasons of FM, then its not significant enough of a result to include anyway.
In case you're wondering, keeping the affiliates for Man City reduces median PA by ~15-20, and I deliberately removed them because I know affiliates mess with the newgen intake results as they can replace one or more of them with one from an affiliate club. I wanted to get a pure English club result.
I noticed that Man City often gets some purely Scotland/Ireland only players, and that once their top player (194 PA) was Ireland only. This is notable because Ireland has low enough of a nation youth rating that a 194 PA Ireland player is pretty highly unlikely. It gave credence to the idea that newgens are tied to local region in some way, so I thought I'd re-examine that aspect too (I'd previously tested 'local region' and found it had zero effect on CA/PA).
Hypothesis: Youth facilities have insignificant effect at a single club, but a nations or divisions teams YF contributes cumulatively to a pool of PA that clubs then draw upon.
Conclusion: False.
Hypothesis: Youth facilities don't effect median PA, but they effect peak PA or the skew.
Conclusion: False.
Although YF1 always eventually produced a high peak PA, I did notice that YF1 alone had a long run of ~155-160 peak PA. This wasn't the case in other YF1 tests, and YF20 produced peak PAs all over the place from 159 to 194. Skew is difficult to definitively conclude on, but I didn't get the impression looking through the results that youth facilities effected the skew.
Hypothesis: Youth facilities contribute to a division pool only.
Conclusion: False.
Hypothesis: Youth facilities contribute to a 'local region' pool only.
Conclusion: False.
Hypothesis: Junior coaching/Training Facilities/Youth Importance of other clubs has an effect somehow.
Conclusion: False.
Hypothesis: Club reputation of other clubs has an effect somehow.
Conclusion: Tentatively, yes.
This was a somewhat surprising finding. From memory, I did find before that club rep had a ~10% CA effect, but not on PA. In this case, we can see that ClubRep1 reduced median PA by ~7.7% alone, and ~12.2% when combined with other low club attributes. I took extra samples to be sure and got the same result.
The strange thing is that ClubRep1 result is normal when combined with YF1. I think there's a clue as to what's going on in the fact that the low median PA results only happened when *all* other England teams had low club rep. But that still doesn't make sense of it.
Hypothesis: Corporate facilities of other clubs has an effect somehow.
Conclusion: Unclear.
Corp 1 did produce lower median PAs consistently, but the difference of ~3-7 may not be statistically significant. Consider that the typical variance of a club each year is ~5 median PA, even after stripping away the outliers.
Hypothesis: Starting CA affects PA, say that +20 CA due to factor contributes +20 PA as well.
Conclusion: False.
Long throws must have increased from 1 lot of 'set piece routines' in my 'mixed' screenshot. Even 'match practice' doesn't seem to do it, as long throws is not a highlighted role attribute. Sessions do seem to train the attributes they say they do. My theory was that sessions with low priority for outfield players, like set pieces, have greater overall growth (which turned out to be wrong).
SaMaHaJoGu said: If you do further testing for set pieces, please do help fill the rest of us in.
I'm making another post here soon that will say my conclusion about set piece attributes (and surmising everything else), which is that they're pretty much useless and it doesn't matter even if your best takers have only say '6' attribute.
For example:
A season may have ~10 penalties. The difference between '6' and '16' penalty taking is perhaps ~30% success rate, so 3 goals total. Top teams probably give away less penalties due to better 'decisions', so realistically this may be just 1 key goal over the season - which may be 1 key game win. However this would no doubt be outweighed by the CA cost or rareness of the high 'penalty taking' player. Only ~30% of good striker options age 20 or under have 10 or higher penalty taking. Therefore it should be considered a luxury and a tie-breaker between 2 equal ~180+ PA players.
HarvestGreen22's data shows us that +6 pace = +40 goals. We can guess +10 penalty taking = +3 goals. So +10 penalty taking = +0.45 pace. So you're better off glancing at the pace/acc/dribbling stats and choosing the one that has 1 extra pace/acc/dribbling.
The Zippo data also suggestively shows corners and free kicks only giving +1 season goal each from +12 each.
Two other ways it can be expressed is 1 consistency = 1.25 set piece = 3 important matches, but this is just a fun piece of trivia not useful for deciding between players.
If you do want a good set piece taker for comfort, best to buy one instead of training up one.
So I came up with some theories.
The first was that perhaps instead of set pieces & technicals in general being squeezed out by low CA, it was instead the 'general' sessions being usually 80-100% priority while the other sessions are 20-60% priority - perhaps the the remaining 40-80% was directed to a hidden pool of attribute growth, just like how 'rest' increases physicals by default. This hidden attribute growth would be significantly more overall growth, which makes sense, as in games focusing on specific attributes often carries an efficiency cost as a game balance.
Perhaps 'priority' can be stacked up to 100%, but anything more simply changes the proportions of that 100%, and each extra session is an extra session of hidden attribute growth. So for instance, 'attacking' (80%) + 'overall' (100%) = 2 x hidden growth and 100% directed growth, whereas 'goalkeeping' (20%) + 'penalties' (20%) + 'outfield' (60%) = 3 x hidden growth and 100% directed growth. This would make sense from what we know SI does, where they use misdirection to keep the true mechanics unknown - in this case, the best training regime would be what appears the least likely and simplistic.
But I did some rough tests, and can rule all of the above out. HarvestGreen22's Quick + Attack + Rest + Quick focus produced significantly better results with less injuries.
I thought I'd share some screenshots, as this shows what you can expect in reality as opposed to a perfect environment averaged 1000 times - one season, England Div 3, 3-4 star coaches, a generous amount of good but not amazing youngsters actually training at the club, just let the chips fall where they may in terms match experience, injuries and so forth (but I'm still selecting amongst the 'survivors' to show here).
Walters_quickattack = Quick + Attack + Rest + Quick focus. 40 injuries.
Walters_mixed = My dense but optimized regime based on theories above. 80 injuries.
Walters_extreme = Complete overload on physicals training + 2 x match practice & 'overall' to boost technicals/mentals. 207 injuries.
So in typical play is the HarvestGreen22 training significantly better? Yes, it gives 1-2 extra Pace/Acc in 1 season, halves the injury rate, and might even have an edge in mentals over a more 'balanced' regime