Awesome comments, thanks for doing this. It's exactly the kind of discussion I was hoping we will have.

Let me first add a few more scenarios just to broaden the horizon a little bit:

------

Scenario C:
A big GPU farm gets to mine 1000 of the 1440 blocks. The remaining 440 blocks are mined by hundreds of servicenode operators each operating just 1 GPU, and small/midsize miners who are on our side!
The big GPU farm is greedy, doesn't understand economics, doesn't listen to hundreds of very vocal servicenode operators (and other community members) who do a lot of persuading (in more inventive ways than we can know right now).
So the big GPU farm votes "zero" all the time. (driving the price down, harming itself out of existence until it has to switch to another coin while us little guys take back control again, ... but I digress!   )

So everybody gets to observe the misconduct of the big GPU farm very closely, and so the small GPU miners (most of them servicenode operators) decide that it's an emergency so they all apply the maximum counteraction they are allowed to: vote "100".

The project might be in a state where the whole community (devs, miners, node operators, etc..) has agreed to publicly recommend a percentage of 20% (based on state of development, functionality and usecases, etc...) .
So interestingly a party can always give a vote that is much higher than the current recommendation, EXACTLY so they can counteract an "unfair" percentage if necessary!

1000 "0" votes averaged with 440 "100" votes is still a very healthy 30.5 %. So the voting collective would have managed to even surpass the proposed goal during this emergency (suicidal GPU farm).

For the same reason we can imagine the opposite scenario where servicenode operators somehow manage to largely dominate mining (which wouldn't be a bad thing if you think about it... it would be like they pay themselves lol) ... but analogous to the previous case miners who think the percentage is too high, can always take drastic countermeasures.

------

Scenario D:
The network has matured and we see an interesting organic behaviour were people have learned how voting works and how the community can act as a whole to decide the percentage.
The next evolution of this would be to use this voting system in case it turns out something is malfunctioning in the current version of the very very complex servicenode network.

Why should miners be forced to continue to pay into a temporarily broken system?
So the community can now react very swiftly, spread the message and reduce the percentage accordingly in just a few hours.
The devs work out what caused the bug, and create an update.
While they are doing this the miners don't feel like jackasses but instead like empowered members of an everimproving system!
After the servicenode network has been fixed, miners return to the last known "good" percentage.
It's always very easy to return to a percentage, everybody just needs to type in the same vote for a period of 24 hours.
(ofcourse I'm assuming here a system of matured miners that understand what is beneficial for the price of SPR)

------

Scenario E:
The network has reached godmode level (the final level) and people don't even need a recommendation anymore to figure out the best percentage.
It has become an inexplicable entangled feedback loop (kinda like a giant orgy) between minining servicenodes and serving miners.
The perfect free market!