( edited 1 day, 23 hours ago by GeorgeFloydOverdosed )
Share this post#1
Link to the post:
So I've known for the better part of a decade now that there's a significant hidden mechanic when it comes to newgens, that seems to be the unique nation itself or something along these lines.
The picture I had about newgen mechanics is:
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
How I found this is by swapping England and Romania completely, after realizing the numbers didn't add up after deducing the relevant factors and trying to calculate outcomes with them. It took a lot of work so I've never repeated it since.
It occurred to me that there's a simpler way of going about it: Make every club, division, and nation have identical relevant attributes, and measure the median PA differences that result.
The idea worked:
France 120.6 England 116.5 Spain 110.3 Germany 110.2 Brazil 109.5 South Africa 107.4 Argentina 106.7 Russia 104.5 Italy 100.8 Israel 98.6 Switzerland 98.4 Singapore 73.0 Gibraltar 64.5
This means that all things being equal, a newgen at a France Div 1 club will have 20% more PA than a newgen at Italian club, on average.
The formula I've come up with based on this is as follows:
((Junior Coaching x 1.5)+(Youth Recruitment)+(Youth Rating/10)+(Factor X))x1.15
Factor X is the hidden nation/division factor, and it is 62 for England with all other nations to be set proportional to it (i.e. Singapore 73 PA > England 116.5 PA = Singapore 25 Factor X > England 62 Factor X).
It's not perfect, but it seems work quite well (keep in mind random fluctuation of ~5 PA is normal with newgen intakes, and up to ~15 PA fluctuation is not unusual):
((10x1.5)+(10)+(13.5)+(62))x1.15 = 115.575 - England (artificial) average 116.5 ((10x1.5)+(10)+(13.5)+(25))x1.15 = 73.025 - Singapore (artificial) average 73.05
((19x1.5)+(20)+(13.5)+(62))x1.15 = 142.6 - Man City (England) 139
West Ham (15JC/15YR) is equivalent to a maxxed out 20JC/20YR club in Italy. A maxxed out 20JC/20YR Singapore club is exactly equivalent to 1JC/1YR in the English Premier League.
Brentford (7JC/7YR) = 106.95 10JC/10YR Argentina = 106.57 15JC/15YR Switzerland = 107.05 20JC/20YR Singapore = 90.27 1JC/1YR England = 89.7
Rough data for other countries (-/+20):
Norway 112.5 China 111 Hungary 111 Japan 108.8 South Korea 107 Mexico 106 Chile 105.5 Australia 104 Bulgaria 103 Finland 102.5 Scotland 101.5 Romania 100.5 Slovenia 100 Wales 97.5 Turkey 93 Belarus 90 Ireland 83.5 Canada 82.5 Hong Kong 79 N. Ireland 67
Here's what was controlled (made identical):
Nation region Nation game importance Nation state of development Nation youth rating Nation reputation Nation tactical attributes (including preferred formations) Nation economic power Nation FA financial power Nation ranking points Club reputation Club morale Club training Club youth coaching Club youth facilities Club youth recruitment Club youth importance Club corporate facilities Club preferred formation Continent regional strength Division reputation
Only Division level 1 teams were altered Professional status is not controlled, I overlooked this one So the controls aren't exhaustive, but they're pretty extensive
Facilities & morale were set to 10. Club rep to 5000. Nation rep to 7000. Nation youth rating to 135. Basically I went for something moderately high, just to help rule out extremes messing things up somehow.
MeanOnSunday said: I thought competition reputation was editable. Also wouldn’t the country population have a large impact? Expand I forgot to add that to the control list, but yes, it was actually controlled for. I'll add it in. I set competition reputation of all Level 1 divisions to 130.
I don't have competition reputation on my list of factors, which means it shouldn't have any effect on PA. I can't actually remember this, since I tested it back in 2020, but I can confirm my 2020 notes say it has no effect on PA. Note that CA works a bit differently.
Country population has no impact - we know it can't be a hidden factor, because there are obvious counter-examples and rationales we can infer: China and India have much higher populations than England. If population mattered, even a little in regards to certain cases, it would make a real mess of things.
Further suggestive evidence for this is that cities and regions don't matter, and turn out to be entirely cosmetic under careful examination.
All the talk you've heard from SI staff about it taking into account economic factor, youth facilities, players being recruited from nearby cities, etc. is baloney.
GeorgeFloydOverdosed said: All the talk you've heard from SI staff about it taking into account economic factor, youth facilities, players being recruited from nearby cities, etc. is baloney. Expand
The SI answer back in 2023, comes about Dynamic Youth Rating (which was failure) and only; and stated that those factors are remain fixed (e.g Economic Factor, State of Development, FA Financial Power and Game Importance).
Appreciate your efforts to test those things, but for hardcore players is not a secret newgen mechanism. What was a surprice, before 2020; was about Youth Facilities - not affecting PA. There was a vast majority of people thought that Youth Players was Training in Youth Facilities, which is not true.
( edited 22 hours, 28 min ago by GeorgeFloydOverdosed )
Share this post#7
Link to the post:
cadoni said: The SI answer back in 2023, comes about Dynamic Youth Rating (which was failure) and only; and stated that those factors are remain fixed (e.g Economic Factor, State of Development, FA Financial Power and Game Importance).
Appreciate your efforts to test those things, but for hardcore players is not a secret newgen mechanism. What was a surprice, before 2020; was about Youth Facilities - not affecting PA. There was a vast majority of people thought that Youth Players was Training in Youth Facilities, which is not true. Expand Not sure what you're saying exactly
If you mean that those 4 factors are what I am identifying, that is not the case.
Of those, only game importance has any effect (at least on PA, but I daresay CA too), and game importance has little effect on median PA. It has a unique effect on the distribution curve, but let's put that aside here.
You make a point of saying these 4 factors are fixed, based on that SI post. I suspect you're taking it to mean that it's fixed, as in it stays the same in spite of whatever value you enter in the editor. This is not true, as game importance can be changed with observable effects. What he would mean by 'fixed' is that it doesn't change once you've started a save.
Also it wouldn't make any sense anyway:
France - 19 economic factor, 17 financial power, developed state, 'very important', 155 youth rating Germany - 19 economic factor, 20 FA financial power, developed state, 'very important', 155 youth rating
How then does France do 10 PA better than Germany on average, when all relevant competition & club factors are identical?
As far as I'm aware, no one else has even suggested this hidden factor I've identified exists, and we see it is significant - the difference between even Italy & England wipes out a 13 difference in Youth Recruitment, judging by the approximate formula.
Yarema said: What are the intake sizes in those countries? Expand I add a manager for each team before intake day, so 16 players per team x the number of teams in the division.
It occurred to me during the process of getting the data that perhaps the hidden factor is merely to do with the number of teams in the division. After all, Singapore & Switzerland & Scotland have fewer teams than the England Premier League, but Italy has 20 and Argentina has 28, so this doesn't explain it.
It's equally valid I think (based on actual past experience) just to take the median of non-managed teams, even if they only retain say 7 of the newgens, but I wanted to make sure + distinguishing between existing & newgen 16 year olds can be painful that way.
Another thing I've noticed is that in spite of all team factors being made identical, club median PAs still tend to follow the conventional reputations. For instance, Maccabi Tel-Aviv is well ahead of other Israeli teams, and same for CSKA Moscow in Russia. AS Monaco is 1st in France, AC Milan in Italy.
I thought initially it's just that these clubs are the ones most likely improve their facilities/rep in those few months, or that perhaps continental competition qualification plays some small role. But then I saw this also happened for a short ~4 month wait until intake, and with nations that don't really have continental competition/qualification of note.
My hunch is that each team actually has a set target PA as part of the hidden factor. I know that sound a bit far-fetched or convenient, but consider the following:
1) From this guide for Italian volunteer researchers for FM20, they are instructed to match a list of average CAs for each team & division:
So it's evident the SI team are already going to these lengths even before the volunteer researcher data comes in. Rather than a general nation or division factor, it could simply be these values are being given a weighting in the outcome.
2) This is speculative, but it makes sense to me from a game balance and narrative perspective why they would include this in the formula. This way, if Liverpool somehow bomb out a few years into the game, they are somewhat cushioned from the prospect of relegation, and even if they are relegated, it allows for them to bounce back. Or if they neglect their Youth Recruitment which we know is very competitive in England and drops with time without investment, then this helps to keep them propped up as a source of quality youth.
Now separate to all that, there are two other things that are relevant to my formula but not included in the formula:
1) Game Importance. This is a difficult factor to measure because it seemingly affects distribution more than median PA, and it also seems to have an effect that reveals itself over several seasons rather than just one (which may carry a more pronounced effect on median PA as time goes on). Thankfully this isn't a factor you can improve in-game, and most of the best nations have the highest game importance anyway, so it's not a huge deal if we set it aside and simply say only expect good newgens from 'very important' nations if you want to be on the safe side.
2) In this thread I detail a curious impact I found in regards mainly to Club reputation and Youth facilities of other clubs within the nation. Club Rep 1 of all other clubs results in ~10% less median PA, while combining it with Youth Facilities 1 of all other clubs somehow cures it completely. Ordinarily I have found that club rep has no effect on PA, aside from Youth Recruitment tie-breaking. The underlying logic to all this remains unclear. Overall, I felt it's not necessary to complicate the formula with this. The point of having a formula would be to quickly deduce things with minimal inputs, and if its ~90% accurate, I reckon that's good enough.
( edited 3 hours, 38 min ago by GeorgeFloydOverdosed )
Share this post#11
Link to the post:
I had a thinkle about my team PA list hypothesis.
It would have to be that it's set for division or nation, otherwise a division 7 club could never improve to a high level. But then how would it determine the ranking of specific clubs, as we see in the actual results?
I realized what instead it could be is the position the club is at in the league at the time of intake. I've done some tests with Luton so far that seem to show this is the case:
Luton 19th/20th = 124, 111.5, 115, 109.5 = 115 average Luton 12th = 122, 107 = 114.5 average Luton 5th = 143.5 Luton 4th = 129.5, 133, 125 = 129.16 average Luton 2nd = 135.5, 135 = 135.25 average
Note this does not mean there is not a hidden factor, as the 1st team in Gibraltar is still worse than the 20th team in the EPL, when all relevant visible factors are equal. But it does explain another piece of the puzzle that can be added to the factors list.
This is still compatible with the team PA list hypothesis in this way: Rankings (i.e. 3rd, 4th, 5th, etc.) in a division could be assigned PAs, and who ranks where is determined on intake day by the actual competition data.
Something else I noticed is that the peak PA also seemed to change, from ~130s for 20th to ~150s for 2nd. What stands out to me about it actually is that the PA range always seemed quite compressed. Usually with very high median PA teams one should get the random 178 or 192 PA newgen. But not here with Luton, where the peak is consistently quite low - 158 or 159 or even 148. Makes me suspect that factors such as Youth Recruitment and perhaps Youth Rating are solely responsible for those very high PA newgens.
Edit: I do not find the same for Balestier in Singapore:
This doesn't mean the Luton case is invalid, but it means that we cannot assume other divisions have a similar position effect. We can say it's true for the English Premier League. It's possible that Singapore simply has identical PAs set for its ranking list, but that is just one of a whole bunch of possibilities, so we cannot assume that.
So I've known for the better part of a decade now that there's a significant hidden mechanic when it comes to newgens, that seems to be the unique nation itself or something along these lines.
The picture I had about newgen mechanics is:
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
How I found this is by swapping England and Romania completely, after realizing the numbers didn't add up after deducing the relevant factors and trying to calculate outcomes with them. It took a lot of work so I've never repeated it since.
It occurred to me that there's a simpler way of going about it: Make every club, division, and nation have identical relevant attributes, and measure the median PA differences that result.
The idea worked:
France 120.6
England 116.5
Spain 110.3
Germany 110.2
Brazil 109.5
South Africa 107.4
Argentina 106.7
Russia 104.5
Italy 100.8
Israel 98.6
Switzerland 98.4
Singapore 73.0
Gibraltar 64.5
This means that all things being equal, a newgen at a France Div 1 club will have 20% more PA than a newgen at Italian club, on average.
The formula I've come up with based on this is as follows:
((Junior Coaching x 1.5)+(Youth Recruitment)+(Youth Rating/10)+(Factor X))x1.15
Factor X is the hidden nation/division factor, and it is 62 for England with all other nations to be set proportional to it (i.e. Singapore 73 PA > England 116.5 PA = Singapore 25 Factor X > England 62 Factor X).
It's not perfect, but it seems work quite well (keep in mind random fluctuation of ~5 PA is normal with newgen intakes, and up to ~15 PA fluctuation is not unusual):
((10x1.5)+(10)+(13.5)+(62))x1.15 = 115.575 - England (artificial) average 116.5
((10x1.5)+(10)+(13.5)+(25))x1.15 = 73.025 - Singapore (artificial) average 73.05
((19x1.5)+(20)+(13.5)+(62))x1.15 = 142.6 - Man City (England) 139
((15x1.5)+(19)+(14.4)+(49.14))x1.15 = 120.796 - Blu-neri (Italy) 120
((14x1.5)+(14)+(14.4)+(49.14))x1.15 = 113.32 - Torino (Italy) 108
((3x1.5)+(15)+(3.5)+(25))x1.15 = 55.2 - Tampines (Singapore) 54.5, 51
((2x1.5)+(9)+(3.5)+(25))x1.15 = 46.575 Balestier (Singapore) 24, 28
((7x1.5)+(8)+(9)+(46.59))x1.15 = 85.2 Yverdon (Switzerland) 70, 90
((17x1.5)+(18)+(9)+(46.59))x1.15 = 113.95 Zurich (Switzerland) 83.5, 111.5
((11x1.5)+(6)+(9)+(46.59))x1.15 = 89.8 - Stade Lausanne-Ouchy (Switzerland) 90.5, 96
((13x1.5)+(13)+(16.3)+(56.05))x1.15 = 120.57 - AMG (Brazil) 126
((11x1.5)+(9)+(16.3)+(56.05))x1.15 = 112.53 - CUI (Brazil) 103
((14x1.5)+(18)+(16.3)+(56.05))x1.15 = 128.05 - FLU (Brazil) 121
((15x1.5)+(17)+(16.3)+(56.05))x1.15 = 128.63 - SEP (Brazil) 130.5
((7x1.5)+(12)+(16.3)+(56.05))x1.15 = 109.08 - FOR (Brazil) 100
((15x1.5)+(14)+(14)+(53.67))x1.15 = 119.795 - Argentinos Jrs. (Argentina) 116
((9x1.5)+(13)+(14)+(53.67))x1.15 = 108.29 - Atl Tucuman (Argentina) 105
((7x1.5)+(10)+(14)+(53.67))x1.15 = 101.395 - Central Cordoba (SdE) (Argentina) 104
((14x1.5)+(14)+(14)+(53.67))x1.15 = 118.07 - Estudiantes (LP) (Argentina) 124.5
((4x1.5)+(11)+(14)+(53.67))x1.15 = 97.37 - Barracas Central (Argentina) 101.5
((19x1.5)+(20)+(14.5)+(56.71))x1.15 = 137.66 - Barcelona (Spain) 126, 138.5
((14x1.5)+(15)+(14.5)+(56.71))x1.15 = 123.29 - Real Hispalis (Spain) 125, 124
((11x1.5)+(13)+(14.5)+(56.71))x1.15 = 115.816 - Almeria (Spain) 112.5, 101
West Ham (15JC/15YR) is equivalent to a maxxed out 20JC/20YR club in Italy.
A maxxed out 20JC/20YR Singapore club is exactly equivalent to 1JC/1YR in the English Premier League.
Brentford (7JC/7YR) = 106.95
10JC/10YR Argentina = 106.57
15JC/15YR Switzerland = 107.05
20JC/20YR Singapore = 90.27
1JC/1YR England = 89.7
Rough data for other countries (-/+20):
Norway 112.5
China 111
Hungary 111
Japan 108.8
South Korea 107
Mexico 106
Chile 105.5
Australia 104
Bulgaria 103
Finland 102.5
Scotland 101.5
Romania 100.5
Slovenia 100
Wales 97.5
Turkey 93
Belarus 90
Ireland 83.5
Canada 82.5
Hong Kong 79
N. Ireland 67
Here's what was controlled (made identical):
Nation region
Nation game importance
Nation state of development
Nation youth rating
Nation reputation
Nation tactical attributes (including preferred formations)
Nation economic power
Nation FA financial power
Nation ranking points
Club reputation
Club morale
Club training
Club youth coaching
Club youth facilities
Club youth recruitment
Club youth importance
Club corporate facilities
Club preferred formation
Continent regional strength
Division reputation
Only Division level 1 teams were altered
Professional status is not controlled, I overlooked this one
So the controls aren't exhaustive, but they're pretty extensive
Facilities & morale were set to 10. Club rep to 5000. Nation rep to 7000. Nation youth rating to 135. Basically I went for something moderately high, just to help rule out extremes messing things up somehow.
Reserved.
Reserved.
I thought competition reputation was editable. Also wouldn’t the country population have a large impact?
MeanOnSunday said: I thought competition reputation was editable. Also wouldn’t the country population have a large impact?
I forgot to add that to the control list, but yes, it was actually controlled for. I'll add it in. I set competition reputation of all Level 1 divisions to 130.
I don't have competition reputation on my list of factors, which means it shouldn't have any effect on PA. I can't actually remember this, since I tested it back in 2020, but I can confirm my 2020 notes say it has no effect on PA. Note that CA works a bit differently.
Country population has no impact - we know it can't be a hidden factor, because there are obvious counter-examples and rationales we can infer: China and India have much higher populations than England. If population mattered, even a little in regards to certain cases, it would make a real mess of things.
Further suggestive evidence for this is that cities and regions don't matter, and turn out to be entirely cosmetic under careful examination.
All the talk you've heard from SI staff about it taking into account economic factor, youth facilities, players being recruited from nearby cities, etc. is baloney.
GeorgeFloydOverdosed said: All the talk you've heard from SI staff about it taking into account economic factor, youth facilities, players being recruited from nearby cities, etc. is baloney.
The SI answer back in 2023, comes about Dynamic Youth Rating (which was failure) and only; and stated that those factors are remain fixed (e.g Economic Factor, State of Development, FA Financial Power and Game Importance).
Appreciate your efforts to test those things, but for hardcore players is not a secret newgen mechanism. What was a surprice, before 2020; was about Youth Facilities - not affecting PA. There was a vast majority of people thought that Youth Players was Training in Youth Facilities, which is not true.
cadoni said: The SI answer back in 2023, comes about Dynamic Youth Rating (which was failure) and only; and stated that those factors are remain fixed (e.g Economic Factor, State of Development, FA Financial Power and Game Importance).
Appreciate your efforts to test those things, but for hardcore players is not a secret newgen mechanism. What was a surprice, before 2020; was about Youth Facilities - not affecting PA. There was a vast majority of people thought that Youth Players was Training in Youth Facilities, which is not true.
Not sure what you're saying exactly
If you mean that those 4 factors are what I am identifying, that is not the case.
Of those, only game importance has any effect (at least on PA, but I daresay CA too), and game importance has little effect on median PA. It has a unique effect on the distribution curve, but let's put that aside here.
You make a point of saying these 4 factors are fixed, based on that SI post. I suspect you're taking it to mean that it's fixed, as in it stays the same in spite of whatever value you enter in the editor. This is not true, as game importance can be changed with observable effects. What he would mean by 'fixed' is that it doesn't change once you've started a save.
Also it wouldn't make any sense anyway:
France - 19 economic factor, 17 financial power, developed state, 'very important', 155 youth rating
Germany - 19 economic factor, 20 FA financial power, developed state, 'very important', 155 youth rating
How then does France do 10 PA better than Germany on average, when all relevant competition & club factors are identical?
As far as I'm aware, no one else has even suggested this hidden factor I've identified exists, and we see it is significant - the difference between even Italy & England wipes out a 13 difference in Youth Recruitment, judging by the approximate formula.
What are the intake sizes in those countries?
Yarema said: What are the intake sizes in those countries?
I add a manager for each team before intake day, so 16 players per team x the number of teams in the division.
It occurred to me during the process of getting the data that perhaps the hidden factor is merely to do with the number of teams in the division. After all, Singapore & Switzerland & Scotland have fewer teams than the England Premier League, but Italy has 20 and Argentina has 28, so this doesn't explain it.
It's equally valid I think (based on actual past experience) just to take the median of non-managed teams, even if they only retain say 7 of the newgens, but I wanted to make sure + distinguishing between existing & newgen 16 year olds can be painful that way.
Another thing I've noticed is that in spite of all team factors being made identical, club median PAs still tend to follow the conventional reputations. For instance, Maccabi Tel-Aviv is well ahead of other Israeli teams, and same for CSKA Moscow in Russia. AS Monaco is 1st in France, AC Milan in Italy.
I thought initially it's just that these clubs are the ones most likely improve their facilities/rep in those few months, or that perhaps continental competition qualification plays some small role. But then I saw this also happened for a short ~4 month wait until intake, and with nations that don't really have continental competition/qualification of note.
My hunch is that each team actually has a set target PA as part of the hidden factor. I know that sound a bit far-fetched or convenient, but consider the following:
1) From this guide for Italian volunteer researchers for FM20, they are instructed to match a list of average CAs for each team & division:
So it's evident the SI team are already going to these lengths even before the volunteer researcher data comes in. Rather than a general nation or division factor, it could simply be these values are being given a weighting in the outcome.
2) This is speculative, but it makes sense to me from a game balance and narrative perspective why they would include this in the formula. This way, if Liverpool somehow bomb out a few years into the game, they are somewhat cushioned from the prospect of relegation, and even if they are relegated, it allows for them to bounce back. Or if they neglect their Youth Recruitment which we know is very competitive in England and drops with time without investment, then this helps to keep them propped up as a source of quality youth.
Now separate to all that, there are two other things that are relevant to my formula but not included in the formula:
1) Game Importance. This is a difficult factor to measure because it seemingly affects distribution more than median PA, and it also seems to have an effect that reveals itself over several seasons rather than just one (which may carry a more pronounced effect on median PA as time goes on). Thankfully this isn't a factor you can improve in-game, and most of the best nations have the highest game importance anyway, so it's not a huge deal if we set it aside and simply say only expect good newgens from 'very important' nations if you want to be on the safe side.
2) In this thread I detail a curious impact I found in regards mainly to Club reputation and Youth facilities of other clubs within the nation. Club Rep 1 of all other clubs results in ~10% less median PA, while combining it with Youth Facilities 1 of all other clubs somehow cures it completely. Ordinarily I have found that club rep has no effect on PA, aside from Youth Recruitment tie-breaking. The underlying logic to all this remains unclear. Overall, I felt it's not necessary to complicate the formula with this. The point of having a formula would be to quickly deduce things with minimal inputs, and if its ~90% accurate, I reckon that's good enough.
I had a thinkle about my team PA list hypothesis.
It would have to be that it's set for division or nation, otherwise a division 7 club could never improve to a high level. But then how would it determine the ranking of specific clubs, as we see in the actual results?
I realized what instead it could be is the position the club is at in the league at the time of intake. I've done some tests with Luton so far that seem to show this is the case:
Luton 19th/20th = 124, 111.5, 115, 109.5 = 115 average
Luton 12th = 122, 107 = 114.5 average
Luton 5th = 143.5
Luton 4th = 129.5, 133, 125 = 129.16 average
Luton 2nd = 135.5, 135 = 135.25 average
Note this does not mean there is not a hidden factor, as the 1st team in Gibraltar is still worse than the 20th team in the EPL, when all relevant visible factors are equal. But it does explain another piece of the puzzle that can be added to the factors list.
This is still compatible with the team PA list hypothesis in this way: Rankings (i.e. 3rd, 4th, 5th, etc.) in a division could be assigned PAs, and who ranks where is determined on intake day by the actual competition data.
Something else I noticed is that the peak PA also seemed to change, from ~130s for 20th to ~150s for 2nd. What stands out to me about it actually is that the PA range always seemed quite compressed. Usually with very high median PA teams one should get the random 178 or 192 PA newgen. But not here with Luton, where the peak is consistently quite low - 158 or 159 or even 148. Makes me suspect that factors such as Youth Recruitment and perhaps Youth Rating are solely responsible for those very high PA newgens.
Edit: I do not find the same for Balestier in Singapore:
Balestier 1st = 19 (75 peak), 32.5 (84 peak)
Balestier 9th = 31.5 (74 peak), 33 (58 peak)
Balestier artificial 3rd = 80.5 (92 peak)
Balestier artificial 9th = 82 (99 peak)
This doesn't mean the Luton case is invalid, but it means that we cannot assume other divisions have a similar position effect. We can say it's true for the English Premier League. It's possible that Singapore simply has identical PAs set for its ranking list, but that is just one of a whole bunch of possibilities, so we cannot assume that.