Are you trying to troll or are you being serious... Practical? How is it more practical to consider a situational value as opposed to the raw values? That doesn't make any sense...
Sorry, not trying to troll.
Maybe the answer to my questions is that, when stacking, the raw defensive value and the practical value approach each other. I did plug a doubly defended village into the simulator for both 8k:2k and 10k:10K. Again with the 6k:3K:250 attack force, the losses incurred by the 16K:4K village were significantly more than 20% more than the losses for the 20k:20K village.
I don't know why this is, probably has to do with variability of defensive values based upon type of attacking unit.
Here is an oddity that I came across while playing:
attacking force: 6k/3k/250
defending force: 20k/20k
farm space lost (defender): 7304
resources lost (defender): 803k
attacking force: 6k/3k/250
defending force: 20k/20k + 2k HC
farm space lost (defender): 7362
resources lost (defender): 892k
The extra defensive value added by the 2k HC translates into
more losses for the defender, both in terms of farm space and resources. (The attacker, of course, lost the same value in both attacks.)
Add enough HC to the mix and the losses will decrease.
If instead, you consider the 16k/4k defenders, adding any number of swords will decrease the defender's losses.
It seems to me that losses are not always inversely related to defensive power (as one might intuit). If the relationship does not hold in "typical" defending situations, then comparing relative defensive strengths would not carry much practical usefulness.
Of course, it may not be practical to determine what is "typical", leaving us back at the imperfect comparison of raw defensive values.