Piperita
The attribute values sorted by how much they impact results per individual player.



It really shows the value of a good (i.e. highly specialised) goalkeeper, even above outfield physical values. And the dominance of speed for outfield players and why the training meta is the way it is.

Funnily enough, the thing I hated the most about the values in #24 (the importance of position between good and perfect values) is what I love the most about #26: Before the system punished players being good at multiple positions as good was not necessarily good enough. Now with IP and OOP there is more incentive to "use" CA to make players talented in more than one role.
I played around with the data a bit.

Methodology: I took the data from the attribute test and converted them into a value based upon how much an increase by one would affect the team's points.



I then took this attribute value and multiplied it with how much the respective training schedules raise or decrease said values. So if a schedule increases Speed by 2, it will get the value of 10,66 as ten players are affected, whereas an increase of 1 in Reflexes would only increase the value by 0,66 as I haven't found an exploit to play with two goalkeepers ;)

This I labelled "Effective Growth", i.e. growth that will affect results and not just fluff up numbers. If we divide the CA growth from this effective growth, we also get an efficiency value -- how much does an increase in CA generate in value.

A few outstanding results under these metrics were:



I personally use N17 when I have a Wednesday game as it fits quite well into the week; for general use I like my recovery sessions too much as it feels like I get injured less with them. Which is why before running the numbers I played the 3-7 meta training as a 3-5 with two days of super rest. This gave my players plenty of regeneration while still developing quite well. I won the CL and had two games each "free week" and yet never had one player in need of a rest.

Now I mostly play Y19, often as a 4-4 variation, especially after rows of congested weeks. I think it could be beneficial to replace [overall] with [attacking] though as Dribbling is a high-value attribute and you can keep your goalie happy if you sprinkle in an O17 every once in a while.

Funnily enough, the most efficient schedule that does not reduce any relevant attributes? The humble [Defending from the Front]. Alas, I don't know how to set up the training groups to get the most effect from this...

One column is red as I do not know if the goalkeeper raises his physical (and with it the high value Agility) and his mentals at the same level as the outfield players.

-------------------

If we look at the meta training, there are some interesting data points to consider



1) Full rest is amazing for both youth players that need to grow their physicals and for old players that need to retain their effective strength.
2) For players age 23 to 26 the rest schedule takes a considerable dip as apparently players in their prime do in fact "need" proper training and guidance to reach their potential before their growth rate drops considerably.
3) Old players that do train lose a lot of physicals that are not compensated by other increases; while full rest loses CA early, players stay close to their effective prime longer, whereas regular training has a longer strong growth period for young-ish players and keeps transfer values up for prime-players.

I personally take the following approach:

1) My main team trains the best schedules, for me N17 and Y19. Most players have room to grow and it gives good gains.
2) My youth team is full of high potential players. Unless they severely lack physicals (then they train with the reserves), they too play the good schedules.
3) My reserve team is mostly a holding pen for less talented youth players and some less-balanced older players with CA left to spare. They train the full rest schedule as it gets the youth players to a good base to be valuable team members and creates specialists out of the older members that, just as a bonus, might have good personalities to mentor the young ones.

----------------

Note: For the raw values for the meta training comparison, there is a minor error in the efficiency: A negative "Effective Growth" should result in a negative "per CA" value whereas a negative CA with positive effective growth should yield high positive effects. This means at 25 years the Rest schedule is the most insanely effective as it still has a great effective growth while decreasing CA. And that at 35 years old players should not even bother showing up for training...