Yarema said: Interesting that league reputation has no effect on growth. I guess in testing scenario that could be the case but I'm pretty sure that CA is limited in a league with bad reputation - for example a 200 PA player might only be able to grow up to 150 CA in a mediocre league. Expand
I've heard the same being said about facilities (better facilities are required for higher CA ranges), but I don't have any evidence to back it up.
BulldozerJokic said: MP has professionalism set to 20 but it's hard to tell what his other hidden attributes are, whereas MC has all of his hidden attributes between 14-20 afair. HoYD slightly affects hidden attributes of your youth intake. In order to minmax their growth, you want HoYD to be MC, or perhaps idealist (this personality has high professionalism and ambition, but othe attributes moght be low and it's not clear how it affects players efficiency on the pitch). Expand
MP with Unflappable and Reserved is probably the best (20 prof, 15-20 pressure, 15-20 temperament) but MC is safer. Spirited with Reserved is also good (15-17 prof, 15-20 pressure, 15-20 temp). AFAIK pressure is arguably the 2nd most important HA with how it affects on-pitch performance.
I’m looking at implemented Q10 as it seems to distribute the training more evenly.
What is [all attack group but CBs]? Is that just Attack Team session? Expand
Your training is split into 3 groups: goalkeeping, defending, attacking. Defenders and DMs are put in the defending group by default so you have to move the FBs and DMs to the attacking group.
harvestgreen22 said: You are right. I saw it in that same video's Comments section, A community player said it. He pointed out that growth is a roll of the dice, with a random chance of full growth or no growth. In terms of numbers, he thinks it's 0.25,0.5,0.75, corresponding to different green arrows.
I set the testing time as 1 season in the hope that this length can reduce the statistical error roughly. Then, for each test, there are 11 players, so the sample size is also increased. In the table, I usually test 3-8 seasons to further reduce the random error Expand
Maybe I misread the post but the arrows denote the amount of growth, not probability. But growth occurs in increments of 1/200 (i.e. 0.1 attribute points) so it will not necessarily increment the shown attribute. The thresholds are roughly what you listed for the arrow types.
Wow, baffling result! Removing [Quickness] from M10 somehow removed all non-physical development that should've come from the Match Practice and Attacking sessions and in turn increased the physical development.
These results indicate that the "weight" hypothesis is incorrect, but additive training also doesn't fit and "more impact of additional focus" only partially fits. I have no immediate explanation for how these results can come about.
On a positive note J11 seems to yield even better low-CA results than a pure resting schedule.
To begin with. Since we already have tests of those with [Quickness] included we can compare to the old results.
E: On second thought maybe valuable information could ve gained from 1 and 2 after all, but only drop the focus and keep [Double Intensity] to see if it 7s the focus that make Rest/Recovery sessions so effective.
I.E: [Attacking Shadow Play]+[Double Intensity] And [Match Practice]+[Attacking]+7x [Recovery]+[Double Intensity]
You need to test these two, right? Do they need +[Addtional Focus Quickness]+[Double Intensity]? Expand
Yes, I'm sorry. I first assumed it was implied but I edited for clarity right as you posted.
Basically I want to compare the two most simple meta programs to see how they react without the Quickness session. I.e. compared to how they performed with a Quickness session.
I'm not entirely convinced that the "assignment according to weights based on sessions performed" hypothesis is proven. I think alternative hypotheses are that they either are very effective physical training sessions or just intensely work the individual focus which ends up being more effective than a physical training session. Looking into EBFM's data which I only had a vague recollection of the Recovery session didn't have as significant results as I recalled and should give higher growth in Jumping/Strength than Pace/Acceleration if it was the case it was as simple as a session that added attributes, so just being effective sessions may not be true either. Rest wasn't included in the data but I would presume it to be very similar to Recovery. His data also suggests Physical sessions being almost as good for Pace, Acceleration, Agility and Balance as Quickness, but with significantly higher impact on Jumping, Strength, Stamina and Work Rate, so further testing where [Quickness] is replaced by [Physical] seems merited.
To test the weighting hypothesis I'd like to see results from one/several of the most effective programs with the Quickness session excluded (E.g. Attacking Shadow Play+7x Recovery, or Match Practice+Attacking+7x Recovery) and compare the results to the programs with quickness included. If the weighting hypothesis is true then excluding the Quickness session would tank physical development as it wouldn't have a weight outside of the individual focus, but I have a feeling that excluding the Quickness sessions will have either a neglible or even positive impact.
Singularity said: I'm doing the same sort of thing as you @Eppstar. All rest until 18 and then switch to V7 at 18
I am getting very large boosts to Pace/Acc too - but I'm getting several guys who get boosts in what I'm thinking of as "Overspill" attributes - freekicks, corners, penalties and long shots
This is about 3 seasons of development
I'm wondering if a 2nd match pracise rather than the attacking would be better Expand
Only Long Shots costs CA of those attributes (and even then it's pretty cheap IIRC) so I wouldn't worry too much about that.
If I had the time, resources and know-how to do these tests I would look into granularity and diminishing/increasing returns on the key attributes as well as matchups in certain roles. I recall at least one test on an older version of FM that highlighted strength as a significant attribute for strikers specifically (and inductively for centre backs as well), it doesn't rate highly on a team level but is it important for box players specifically?
Same goes for advantage/disadvantage in the relative stats (e.g. most of them). We see the effects of +10/-10 and in the FM Arena tests +5/-5, but we can't assume linear relationships. Looking granularly at increments of 2 could see if it is most important to just have an edge or if the effect becomes significant when you're outclassing your opponent (which is seldom feasible in a save and in such a case you'll win easily regardless of your knowledge of key attributes). I could imagine duel-related attributes like jumping and strength being a case of being ahead is what primarily counts with diminishing returns (and being significantly more important at certain positions as earlier mentioned), whereas pace/acceleration sees you advantage increasing exponentially the higher you go - especially on a team level.
Going deeper how do the valuable attributes interact? We know from tests that the poor attributes can be done without all at once but do the good ones boost each other, so I'm not sure it's worth the effort investigating whether there is a scenario where technique could be slightly beneficial. It's more interesting to see where and how the provably beneficial attributes interact. As mentioned above is 15/15/15 Pace/Acc/Dribbling better than 17/17/10? I could see dribbling having a multiplying effect with a pace advantage compared to equal like in the isolated experiment.
I think one thing that is overlooked is that rest sessions increase physical development.
The "no training" experiments yielded physical beasts but tanked most mental/technical abilities and 200PA players wound up in the 110 CA range. https://m.youtube.com/watch?v=WJynDrBhTUw
Considering the importance of physical attributes you want to maximize their development but not to the point where a player can't reach his PA, so I surmise an optimal schedule contains 2 Physical sessions, 1-2 Resistance and a few Attacking/Defending to make up for the technical/mental decline, perhaps giving slight development in those areas. Filling up the schedule because you can is therefore not necessarily good.
Then you move on to which technicals and mentals you actually want to maintain/develop (dribbling, concentration, anticipation) and which you would happily see decline. E.g. Decisions eats a ton of CA for no verifiable benefit, but unfortunately seems to grow significantly with almost every non-physical session. And while not really verified in the attribute tests, I have seen anecdotal evidence that some technical attributes seems to have an impact where others don't (crossing does seem to yield significantly higher quality crosses based on in game metrics while passing/finishing don't seem to yield significant statistical impact).
From this perspective Possession seems to be a better session than Attacking. It has slightly better physical development, the same impact on dribbling and first touch and a lower increase in decisions. What makes me wary is that while it tanks finishing instead of significantly improving it, it also tanks crossing. If crossing does have a ME impact that would be impactful considering the importance of crosses in FM24. We also have the technical sessions, where some have good attribute profiles such as Transition Press and I don't know if it has been tested but I have seen claims that these also impact match performance in the short term.
I've heard the same being said about facilities (better facilities are required for higher CA ranges), but I don't have any evidence to back it up.
MP with Unflappable and Reserved is probably the best (20 prof, 15-20 pressure, 15-20 temperament) but MC is safer. Spirited with Reserved is also good (15-17 prof, 15-20 pressure, 15-20 temp). AFAIK pressure is arguably the 2nd most important HA with how it affects on-pitch performance.
I’m looking at implemented Q10 as it seems to distribute the training more evenly.
What is [all attack group but CBs]? Is that just Attack Team session?
Your training is split into 3 groups: goalkeeping, defending, attacking. Defenders and DMs are put in the defending group by default so you have to move the FBs and DMs to the attacking group.
I saw it in that same video's Comments section, A community player said it. He pointed out that growth is a roll of the dice, with a random chance of full growth or no growth. In terms of numbers, he thinks it's 0.25,0.5,0.75, corresponding to different green arrows.
I set the testing time as 1 season in the hope that this length can reduce the statistical error roughly. Then, for each test, there are 11 players, so the sample size is also increased. In the table, I usually test 3-8 seasons to further reduce the random error
Maybe I misread the post but the arrows denote the amount of growth, not probability. But growth occurs in increments of 1/200 (i.e. 0.1 attribute points) so it will not necessarily increment the shown attribute. The thresholds are roughly what you listed for the arrow types.
https://pixeldrain.com/u/xQT8i2Bu
Wow, baffling result! Removing [Quickness] from M10 somehow removed all non-physical development that should've come from the Match Practice and Attacking sessions and in turn increased the physical development.
These results indicate that the "weight" hypothesis is incorrect, but additive training also doesn't fit and "more impact of additional focus" only partially fits. I have no immediate explanation for how these results can come about.
On a positive note J11 seems to yield even better low-CA results than a pure resting schedule.
[Attacking Shadow Play]+7x [Recovery]
2.
[Match Practice]+[Attacking]+7x [Recovery]
3.
[Attacking Shadow Play]+7x [Recovery]+[Addtional Focus Quickness]+[Double Intensity]
4.
[Match Practice]+[Attacking]+7x [Recovery]+[Addtional Focus Quickness]+[Double Intensity]
5.
[Physical]
6.
[Quickness]
7.
[Physical]+[Addtional Focus Quickness]+[Double Intensity]
8.
[Quickness]+[Addtional Focus Quickness]+[Double Intensity]
ok , Now 1-4 on list.
Do you need 5-8? I don't test 5-8 if you don't need it
Oh, I don't think tests without double intensity/focus are required.
Just:
[Attacking Shadow Play]+[Additional Focus Quickness]+[Double Intensity]
And
[Match Practice]+[Attacking]+7x [Recovery]+[Additional Focus Quickness]+[Double Intensity]
To begin with. Since we already have tests of those with [Quickness] included we can compare to the old results.
E: On second thought maybe valuable information could ve gained from 1 and 2 after all, but only drop the focus and keep [Double Intensity] to see if it 7s the focus that make Rest/Recovery sessions so effective.
I.E:
[Attacking Shadow Play]+[Double Intensity]
And
[Match Practice]+[Attacking]+7x [Recovery]+[Double Intensity]
Attacking Shadow Play+7x Recovery
2.
Match Practice+Attacking+7x Recovery
You need to test these two, right?
Do they need +[Addtional Focus Quickness]+[Double Intensity]?
Yes, I'm sorry. I first assumed it was implied but I edited for clarity right as you posted.
Basically I want to compare the two most simple meta programs to see how they react without the Quickness session. I.e. compared to how they performed with a Quickness session.
To test the weighting hypothesis I'd like to see results from one/several of the most effective programs with the Quickness session excluded (E.g. Attacking Shadow Play+7x Recovery, or Match Practice+Attacking+7x Recovery) and compare the results to the programs with quickness included. If the weighting hypothesis is true then excluding the Quickness session would tank physical development as it wouldn't have a weight outside of the individual focus, but I have a feeling that excluding the Quickness sessions will have either a neglible or even positive impact.
FM23 training session data (EBFM): https://www.dropbox.com/scl/fi/iil3uzylhhkakvw7a4eex/Training%207%20for%20Dropbox.xlsx?rlkey=8qe1b84rqx7623tzbkxwjts8b&e=4&dl=0
Tl;dr:
Do a comparative test where you drop the Quickness session altogether:
[Attacking Shadow Play]+quickness focus and double intensity
And/or
[Match Practice]+[Attacking]+7x[Recovery]+quickness focus and double intensity
I am getting very large boosts to Pace/Acc too - but I'm getting several guys who get boosts in what I'm thinking of as "Overspill" attributes - freekicks, corners, penalties and long shots
This is about 3 seasons of development
I'm wondering if a 2nd match pracise rather than the attacking would be better
Only Long Shots costs CA of those attributes (and even then it's pretty cheap IIRC) so I wouldn't worry too much about that.
Same goes for advantage/disadvantage in the relative stats (e.g. most of them). We see the effects of +10/-10 and in the FM Arena tests +5/-5, but we can't assume linear relationships. Looking granularly at increments of 2 could see if it is most important to just have an edge or if the effect becomes significant when you're outclassing your opponent (which is seldom feasible in a save and in such a case you'll win easily regardless of your knowledge of key attributes). I could imagine duel-related attributes like jumping and strength being a case of being ahead is what primarily counts with diminishing returns (and being significantly more important at certain positions as earlier mentioned), whereas pace/acceleration sees you advantage increasing exponentially the higher you go - especially on a team level.
Going deeper how do the valuable attributes interact? We know from tests that the poor attributes can be done without all at once but do the good ones boost each other, so I'm not sure it's worth the effort investigating whether there is a scenario where technique could be slightly beneficial. It's more interesting to see where and how the provably beneficial attributes interact. As mentioned above is 15/15/15 Pace/Acc/Dribbling better than 17/17/10? I could see dribbling having a multiplying effect with a pace advantage compared to equal like in the isolated experiment.
The "no training" experiments yielded physical beasts but tanked most mental/technical abilities and 200PA players wound up in the 110 CA range. https://m.youtube.com/watch?v=WJynDrBhTUw
EBFM's tests actually indirectly suggest this as well even though he didn't specifically test the effects of the rest session - but similar non-training sessions such as Recovery, Team Bonding, Match Review etc. yielded significant physical development. https://www.dropbox.com/scl/fi/iil3uzylhhkakvw7a4eex/Training-7-for-Dropbox.xlsx?rlkey=8qe1b84rqx7623tzbkxwjts8b&e=2&dl=0
Considering the importance of physical attributes you want to maximize their development but not to the point where a player can't reach his PA, so I surmise an optimal schedule contains 2 Physical sessions, 1-2 Resistance and a few Attacking/Defending to make up for the technical/mental decline, perhaps giving slight development in those areas. Filling up the schedule because you can is therefore not necessarily good.
Then you move on to which technicals and mentals you actually want to maintain/develop (dribbling, concentration, anticipation) and which you would happily see decline. E.g. Decisions eats a ton of CA for no verifiable benefit, but unfortunately seems to grow significantly with almost every non-physical session. And while not really verified in the attribute tests, I have seen anecdotal evidence that some technical attributes seems to have an impact where others don't (crossing does seem to yield significantly higher quality crosses based on in game metrics while passing/finishing don't seem to yield significant statistical impact).
From this perspective Possession seems to be a better session than Attacking. It has slightly better physical development, the same impact on dribbling and first touch and a lower increase in decisions. What makes me wary is that while it tanks finishing instead of significantly improving it, it also tanks crossing. If crossing does have a ME impact that would be impactful considering the importance of crosses in FM24. We also have the technical sessions, where some have good attribute profiles such as Transition Press and I don't know if it has been tested but I have seen claims that these also impact match performance in the short term.