Lots of good points in this thread, but I wanted to add one that hasn't been said explicitly for a while.
The match-making algorithm doesn't need to defend perfectly against griefing, just well enough to make griefers look elsewhere. For example, say a pair of them repeatedly hyperspace into a system until they end up in the same instance by chance. Then a newbie jumps in - without knowing they're griefers, the match-making system could easily say "he's harmless and they're competent, there's no possible way he could have fun with them so he gets a new instance". But when a deadly bounty hunter turns up, the same algorithm could easily subject the evil pair to the ordeal of a fair fight.
Even if they could exploit the system well enough to find their prey half the time, would it really worth the hassle when they can just go give people a hard time in EVE?
The match-making algorithm doesn't need to defend perfectly against griefing, just well enough to make griefers look elsewhere. For example, say a pair of them repeatedly hyperspace into a system until they end up in the same instance by chance. Then a newbie jumps in - without knowing they're griefers, the match-making system could easily say "he's harmless and they're competent, there's no possible way he could have fun with them so he gets a new instance". But when a deadly bounty hunter turns up, the same algorithm could easily subject the evil pair to the ordeal of a fair fight.
Even if they could exploit the system well enough to find their prey half the time, would it really worth the hassle when they can just go give people a hard time in EVE?