I'll admit I'm becoming more confused about this question of "bromide drag", rather than less confused, the more I read! I'm happy to go with your definition, and whether there is actually bromide involved or not it seems to be a widely-accepted name given to the problem of the streaks adjacent to the sprocket holes, and as you say caused by insufficient agitation. I think another cause can be too-regular agitation (as in continuous consistent agitation) with uneven development caused by developer flow. I have read in earlier threads on this forum that there is no "bromide" as such in the Rodinal chemistry, but whether this is true or not makes no difference to the resulting uneven development.
I was interested in the following statement in the Ed Buffaloe article:
"Bob Schwalberg, who once wrote for Popular Photography, maintained that too much agitation would interfere with adjacency effects, and recommended no more than ten seconds per minute of gentle agitation with Rodinal."
I have seen many people write about the compensating effect of various developers (ie the characteristic of not over-developing the highlights while allowing full development in the shadows), and understand that this usually required reduced agitation. But I was was surprised to see in the statement quoted above that as much as 10 seconds of gentle agitation per minute would not prevent the compensating effect.
So - how much is too much?