Proof that having the sighted revision as the standard view (which you have to for it to be meaningfully useful) doesn't result in a drop in editing rate.
Proof that en would be able to keep up with the required rate of sighting (We have a hard time marking new pages as patrolled at the moment)
If you want to use drug articles as the foundation for your argument proof that it would provide any protection against subtle vandalism.
There is only one way to prove any of that, though, and that's giving it a go. If we only make the flagged rev the default version for articles that are currently (semi-)protected keeping up with the sighting would be easy - that's the version of the proposal that should be completely uncontroversial.