The number of delegates won't be fixed.. Hence why I posted about the n delegate model and how it prevents top down corrupt entities from retaining control within the voting framework. N is determined dynamically in the same way. I believe min is 101 and there is a max number that isn't outrageous.
You are correct (I didn't understand your earlier post), the number of delegates is set by stake holder voting, and I haven't analyzed the effect of that structure. As TPTB says, as these systems get more complex the game theory becomes increasingly impractical to analyze.
At first glance it seems dangerous to me since a large malicious stakeholder now has an additional degree of freedom, namely "pack the court" (or perhaps, depending on the specific voting rules, blocking an increase supported by others). But as I said it is extremely difficulty to analyze.
Anyway, this is irrelevant to the question whether it is proper to consider Bitcoin as "Delegated PoW" just because right now, in 2015, most miners happen to delegate. But honestly is that even true by hash rate?
Of the large "pools" Ant, Bitfury, KnC, 21, and maybe some others are certainly running a lot of
their own mining gear (though these pools may have independent miners too, in unknown amounts). That's concentrated mining power for sure, but it's not really delegated.