This thread is just my theory and is related to key attributes and single role players.
Starting with new training system you are allowed to choose, more or less ,what attributes to train and improve for your players.
Now the question is :
are the key attributes relevant in game engine calculation or just the quality counts?
I am sure that our opinions are divided, and each of us has own felling.
Anyhow, having above starting point I summarized bellow how many key attributes has each single position,
dl/dr - 8
dc - 8
mc - 10
dml/dmr - 8
dmc - 10
amr - 8
mr /ml - 7
amc - 8
st - 9
As you can see, the position with less key attributes is ML/MR(7), that means ML/MR position has the biggest chances to have a big discrepancy between key attributes and non-key attributes, on other side is mc/dmc which has the most key attributes (10).
From my history matches I noticed that playing vs teams with big discrepancy key attributes players were the hardest matches, so I am almost convinced that key attributes are reflected into match engine.
Players with multi role(3) have more key attributes and for long term are not the best option to get big discrepancy in attributes, but they give you the possibility to make changes during game without make substitutions.
My next step is to rebuild my team in such way and buy players with only one role and keep them for 7-8 seasons and try to monitor the performance vs teams with the same quality and why not high quality.
If someone had this in his mind or did it already please share with me the results.