imo, there is no maths of value aside from physical comparisons of the wheels. And that tells you only a part of it. ie roll-over angles don't also consider tyre deformation differences at impact point. Volume variances. Contact patch etc etc. Too complex to model with any validity.
Would you try to balance variables between wheels, or keep tyres and pressures optimised for each wheel as a rider would? How do you test when off-road riding and riders themselves vary so much? I guess a rider, 2 or 3 otherwise near-equal bikes, a track and a HRM can help them decide what works best for them (many mags have done this - a feature rather than a test imo, but still, some result trends seem to appear in them), but I wouldn't call that 'doing the maths' as it can't be applied to riders or terrain equally.
I've done back-to-back runs on 3 wheel sizes and the only thing of value was how I felt the bikes rode differently and of that, what I would attribute to the wheels as one of those 'princess and the pea' kind of riders. I've also done roll down / along types of 'tests' on different tyres and pressures.
After all that, I only had more questions about accepted component designs and what we feel is important on a bike. Rim diameter isn't as important as we may think / in the way the generalisations suggest imo. ie, why all the focus on just that aspect of the wheel? We're happy to challenge one accepted norm yet not others, or we get distracted by one aspect and lose focus. Perhaps because (like suspension) it's too complex to sell easily without being distilled down to easily digested ideas.