There is one very crucial problem with the 3RR, which needs to be taken into consideration by the sysop who is banning. Sometimes particular articles attract several edit warriors, and often it's two or three against one. If that one person "uses up" his three reverts, the others still have several to go before they are also banned. How can we avoid an unequality like this from happening? I otherwise warmly endorse the 3RR.
Jfdwolff