Can you elaborate why you think this is a bug? I'm trying to help, but I need data, not just flames. Something like this (here are your highest paid players' minutes and game shape without the PL minutes):
Ben-Yossef - 67 minutes - strong game shape (nothing wrong here)
Schlechtbach - 65 minutes - respectable game shape (this one *could* be better, but these things happen, you're not guaranteed strong game shape with good minutes)
Gajáry - 57 - strong game shape (again, nothing wrong)
Mizrachi - 59 - strong game shape (same as above)
Espada - 70 - respectable game shape (again, *could* be better, but I don't know what was this players starting GS, as he played 88 minutes two weeks before)
Chicharro - 48 minutes - strong game shape (nothing wrong)
@Radiobasket; I checked the weekly minutes of your players with highest salaries too:
Mühlestein - 71 minutes - proficient game shape (that's ok, right?)
Knobeler - 80 minutes - average game shape (80 minutes definitely lowers your GS)
Lagana - 79 minutes - mediocre game shape (same as above)
Avagliano - 75 minutes - average game shape (can't expect 75 minutes to raise the GS, can you?)
Bartolomei - 62 minutes - respectable game shape (this one would be suspicious if he hadn't played 84 minutes the week before)
Abraão - 69 minutes - strong (again, nothing wrong)
A widely known rule is that in order to preserve his GS, a player needs to play between 48 and 75 minutes (even less if you don't have a doctor with a massage speciality). Looking at all your players, I see only one player with GS respectable (Schlechtbach) which could maybe have better GS, but you must remember that there is a random factor in the calculation of GS. Others have the expected levels of GS considering their minutes. I truly hope people will look at this thread and judge for themselves. There was only a display error (which was promptly fixed), so there is no need for sarcasm (which is extremely rude).
So how is the collapse in game shape justified?
@SpicyMcHaggis™: There is no definite proof that PL minutes were counted, and looking at the code, we determined that it they definitely were not. So, the collapse is a direct consequence of the minutes the players played that week + the minutes from week(s) before + an amount of random.