You ask about returns for skill....
This depends what you want !
If we want to reward pre-planning (having the unit on hand) and make
studying cost effective for more than single person units,
then the diminishing returns for study are bad.
It should be tied to time to attain ... but compared with recruit cost.
I can't seem to come up w/ a formula...
here's some bits of it.
if level 0 produces 'x' in time t
and recruiting costs 'y'
and studyin for time 't' produces leavl 'L' at cost 'c'..
To make it more cost effective to study before working...
In some time 'z' ... the payback period.... level L should produce
at least as much as level 0 working continuously for time t +z,
plus mtce for time t+z
plus studying costs.
It also has to be more cost effective to study than to recruit...
So, if I can recruit R men in time t at cost C
production of R+1 men in time t x(R+1)
minus mtce of R+1 men for time t+z -(t+z)(R+!)(mtce)
minus recruitment cost or R men -C
must be less than
output of the 1 guy who studied...
production in z at level L
minus cost of studying
Z needs to be a reasonable number. Short enough that the skill still
exists at the end of it :-)
I feel that the returns on combat should follow the same curve as everything else, even though they are non monetary.
While there should be a reward for preplanning/having the unit already, it shouldn't be too high, or it is yet another way to freeze out newbies.
Hope this makes sense. I was up late last night, and am quite tired this AM.