The math itself is no mystery to me, but rather the relative values being used. The rules on checks say
"Each leader rating check is essentially the computer generating a Random(x) value where if the result is less than the leader rating then
the check is passed"
This was the source of my uncertainty, since as x increases, the relative importance of leader skill values decreases. What just clarified this, somewhat, for me was the example below the rule (in blue) where it suggests that the value of x starts at 10, and is then modified by penalties for range, CP limits, support squad TOE, etc. It would have been great if the rules themselves were more clear about this, and frankly a modicum of doubt remains as to whether or not x begins as 10 is in fact always true, or at least in what circumstances, aside from penalty application, it is not.
So yes, assuming penalties are nil, utilizing CoC would (best case scenario) reduce failed checks at the FK level by up to 40% (Kesselring is an 8, but check value is doubled for going up CoC).
e.g. Richtofen is also an 8 Air, and would pass an unpenalized skill check with Kesselring directly above him in CoC at a little over 88% (AG & OKH will still factor in, though very small, 1/40 & 1/80 improved fail rates respectively, before factoring in very likely range penalties as well.
Now, this begs the question if really the best solution would be to have Kesselring as AG command, with only subordinate Luftflottes below him (and no usage of FK whatsoever). Of course now the challenge would be to avoid CP penalties on Kesselring, which would make this pointless, as well as affording the AP costs to create such a structure.
Forgive me for going through the examples above that I'm sure you already understand, but a) it helps make sure that you know what I do or don't understand, b) it helps me to put it down in writing and c) as long as I am, maybe some other noob (besides me) will benefit from it.