I think all 3 of these conclusions are reasonable and are similar to theories commonly discussed here.

Sometimes I think that balanced players play better in some formations and in some specific matches because their involvement is a more balanced one. Equally heavily skewed players, say for attacking, can perform exceptionally in some instances because everything demanded of them has been in an attacking sense.

I experimented, and still am, with a team doing 100% defensive counterattacking and all Defenders and Midfielders are balanced for Defence and Physical and perform well, but if I thow in an 'attacking skewed' MC its more unpredictable.

Other theories include some left-footers being better on the right (or right on the left).

I also think there are some players that are destined to be good performers if their role can be found. So I can quite happily go along with theories that there are unexpected good things 'programmed in' that just need to be found.

Finally, yes I both love and hate the unpredictability of some events. I'd guess that the engine has some variables built into it that have such a wide range of values, however unlikely, that a defeat or a win is always possible, given a (un)favourable wind etc.