One of the problems with 3RR-policy enforcement is that admins are supposed to treat equally the editor who is inserting an unreferenced, unsubstantiated claim, and the editor who is trying to get rid of that claim. One is violating [[Wikipedia:Cite sources]] policy, while the other is trying to enforce it. Yet both are blocked.
If the editor trying to enforce policy isn't allowed to violate 3RR, then s/he must go through dispute resolution; put up an RfC (which rarely brings useful results); or apply for mediation (which can take months to arrange). Meanwhile the nonsense sits there for 24 hours; then maybe there's another brief flurry of reverts, then it sits there for another 24 hours; and this can go on for weeks, until the less determined editor backs off.
If the process takes priority over the product, that's fine. But if it's the production of an accurate encylopedia that is the priority, then this is not fine.
Surely, for this reason, when looking at 3RR violations, admins should be allowed to take into account who was violating Wikipedia's editorial content policies and who was trying to preserve them.
Slim