Unsure why you're being downvoted; this is one of the best comments on the thread (and I remember this story from the last time it was posted).
You might say abuse breeds abuse in that someone who for some reason is playing doormat is emitting pheromones for people who like to trample on others.
This is absolutely true. Perhaps some readers were confused by your metaphorical use of 'pheromones' to mean signalling in general. A great example of this is griefing behavior in MMORPGs (and trolling in general, but in games it's already quantified and thus far easier to measure). Most games implement some sort of safe zone and/or NPC policing function to prevent griefers from hassling new players to the point of wrecking the game, which is the griefers' underlying and often unconscious objective (so as to 'own' the territory of the game space even if this is poisonous to the growth of the player pool).
Denied the ability to pick on newbies, griefers then usually collect in small packs and lurk around entry-exit routes to danger zones (whether from NPCs or territorial conflict) where there's a possibility to target outcoming damaged players or incoming ones pushing up against their skill envelope. Griefers like to think of themselves as apex predators, but typically lack the self-discipline and strategic vision required to organize as such, so more often than not they occupy the same environmental as scavengers such as hyenas, vultures etc.
I haven't kept up with the latest research on this, but I recall that EVE had an economist on staff several years ago and I'd imagine that the larger participants in that market are open to or already working with sociologists, game theorists, and other quantitative social scientists to better understand the dynamics of their virtual ecosystem.
For 'nice' players (in games and in life) who don't comfortably slot into large teams, the usual advice is to be more of an asshole. And while that's partly true, being an armored up lone wolf will only take you so far. Unless the system as a whole is dysfunctional, individual lone wolves are never competitive against anything bigger than a small-medium team. However, lone wolves can team up and be very effective; to do so they (obviously) have to overcome significant trust barriers, but can succeed by maintaining smallish flat structures and growing hierarchies below those.