Thank you for your contribution, I have added you to the OP as well as created a new spreadsheet NRG CALC with the correct numbers relating to how many energy per focus at each increment being tested. Also, the above equations are interesting. Can you define the variables for me so I can understand it more completely. More specifically, what does f equal, i equal, nerf equal, and socket equal. The others I understand but some may not so please elaborate.
Thanks HL
PS. Nice name

That was for the beginning a more general idea or type of description, for I am not so sure about the English names and abbrev. in Algebra.
Arithmetic rows are e.g. Function term sum(2n): 2+4+6+8+10+... = 2,6,12,20,30,....
Nerf and socket was just my funny naming for constants, cause that's what they would do in a formula.
Nerf (or boost) would be an overall multiplier.
This might or might not be present (and if not it still is simply 1). More likely it is, given patch comments of nerfing by x%
Socket is a minimum value you get independent of anything else
So one of the possible basic equations (only one degree of freedom / one parameter changing / one variable) is polynomial.
One parameter that clearly seems to be one variable only is ENG cost
ENG = round ( n * ( a* SKL * SKL + b * SKL + c ) ) + s
ENG Energy
SKL Skill level
a,b,c,n,s constants
Round Rounding function ( 1,49999= 1, 1,500 = 2 )
Which is a version of the general polynomial function. Y= ax^2+bx+c
At least this type of formula is well fitted to describe a decreasing or increasing growing curve.
An alternative function is y=a*log(bx)+c
First thing is to establish the a,b,c terms as they fully describe the formula (if the curve is shaped this way at all).
If I'd do the programming, I would use limited groups of those values and do the relative balancing with my n and s constants.
If one does a graph in Excel of the values and adds a trend line of the various types to it, those constants are just what u get. and in addition you get the regression Coefficient R which describes the quality of the fitting.