I am going to use three values as a preliminary ranking/points system - Offense, Defense, and Mobility.
Offense will be the average number of wounds caused per turn (one shooting plus one melee phase) against a 4+ save. 4+ will be used to make rend relevant without some extrapolation, and because (to the best of my knowledge) -3 is the highest rend on any unit, at least as a base stat. I will not be assuming that 1 automatically fails or that 6 automatically succeeds. As this number is quite small compared to the others, it will be multiplied by 10 as a baseline.
Defense will be the average number of wounds the model can take before dying. If the model ignores some amount of rend, for the sake of preliminary numbers, this will be applied as an extra .5 save per point of rend ignored. This is to account for most instances of ignoring rend only ignoring certain rend values rather than reducing it, so while valuable, this cannot be said to be as good as armour save. I am then going to add the unit's bravery reduced by 6, as this seems like a reasonable assumed 'average' for now, as well as being the point at which the unit may lose models for taking a single casualty. Depending on the size of the other values on average, I will either add the bravery value in whole or half as a starting point.
Mobility will be the range the model could move on a turn assuming it took every opportunity to move (run or charge, etc) at an average roll (7, in the case of a normal run or charge). This will be very similar for all units without major movement advantages or disadvantages.
I will be providing three sets of numbers in relation to abilities that would modify the above 'conditionally' (e.g. bonuses from unit size) - one assuming 0% uptime, one assuming 50% uptime, and one assuming 100% uptime.
For now, I will not be addressing spells or other non quantifiable special abilities. I have not entirely worked out how monsters or other models with stats that vary based on number of wounds taken. For now, I will try base statlines only, but will think on this further.