Again though, this board is trying to predict what happens in what will likely be a chaotic environment to start with, and IMO the only way of knowing if this is a better idea than DRK's static mn's is to try this out live.
We will soon be testing this, and if - as I predict - a MN requirement starts at 1 SPR, we don't even need to make the extensive tests on the testnet alone (where only a few people participate).
We can
direcly go live (after preliminary tests on the testnet) because what is there to lose if you can run a MN with such a ridiculously low amount of 1 SPR?
It's not like we will have 1000 Masternodes setup in 1 week.... or ARE WE?

Since they are so cheap? Holy Spreadonia!
The people themselves will decide how secure they think the development version now is, and how much SPR they want to have exposed.
Another thing that SPR could do completely different than DRK.
I could be wrong though, it makes sense to first test everything in a testnet, but that's the thing: you never get a real life situation out of a testnet, so it never is 100% reliable for debugging... darkcoin has shown this many many times)