I think you're discounting the amount of time that can be be sucked up by patrolling. For instance, my longish watchlist has very few of the "popular" articles on it, but still gives me some 500 edits to review each day. Of those, most are by editors I know and trust, so I don't look at those unless the summary line is interesting, but that leaves maybe 100 by anons. Of those, maybe 5 of the anon edits are the obvious "Joey is gay"-type vandalism, but since there is no way to tell which ones they are without bringing up a diff, I need to look at all of those edits. That can easily suck up an hour - very often at least one of the vandalisms is "complex" in that it involves multiple edits, and maybe a ham-handed incomplete attempt to fix, thus requiring careful study of the history to make sure all is scrubbed.
I think the solution to this is to come up with ways to streamline this process. What if watchlists could only show anon edits (or better yet, include anon edits, edits by certain users you designate, and edits by users with fewer than X edits)? What if there were ways for groups of users to share a watchlist and check off items as they complete them (this could be done for any ad-hoc group of users)? What if we could attach references directly to pieces of text, or maybe even better, to individual edits? What if we created a project to systematically go through all the edits made by IP addresses? What if we added to that a project to go through all the edits made by users who contributed fewer than say 10 edits? Time is being wasted because vandalism hunting is way too ad-hoc.
Blocking is an inherently flawed solution to vandalism unless you suggest that we lock down the wiki to essentially just the admins. Good idea or not, that's not going to happen, the history of Wikipedia has brought together too many people who would vigourously oppose it, so if you want a locked down "wiki" I suggest you start a fork.
Sometimes blocking is an adequete temporary solution for the times when a longer term fix is still in the works. But as Wikipedia grows and vandals become more sophisticated, blocking is unlikely to be a good solution.
I think we actually block far too much already, especially with user blocks. I'd suggest that most blocks actually tend to make it harder to find and revert vandalism.
Of the other 95 anon edits, most are trivial - spellfix, commas, random
rearrangement changing good English into broken English :-), etc. This leaves a handful of valuable edits, but the average total is less than I could have added in the same hour just working from the books in my personal library. Nevertheless, I do the patrolling because it seems that many of the pages I'm watching have no other reviewers - more than a few times I've overlooked an anon's trash and it went unnoticed for days or weeks.
I think a study of anon edits would show that the vast majority of them are trivial, and the remaining ones are by people who would have created an account anyway. What can't be as easily measured, and what I suspect is the case, is that there are a lot of users who initially edited the wiki anonymously and later got hooked and created an account.
So yes, we're keeping the random vandalism under control, but IMHO
just barely, and at the price of time that should be going into development of better content. I think we really need to consider whether unlimited anon editing is helping or hurting our primary goal of encyclopedia writing.
Stan