I still think what Microsoft did could work. Their new Flight sim game, (I don't play games, I just read about this), was being leaked, cracked, whatever you call it, and put on these torrent sites. Well, MS, I don't remember, either started to DL it themself, or leaked it themself, but either way, they started to resolve all IP's connected to said torrent, and calling ISP's of the people who were downloading. In other words, they'd call say, Comcast, and say, IP 127.0.0.1 (example) at 9:26pm standard time, 3/25/07, was illegaly downloading cracked versions of our software. The ISP in turn was calling the homes of these people, and telling them that their activity was being monitored, and that they will not tolerate illegal activity, and gave them a "we'll stop your service" warning.
Now, of course you could use proxies, but the type of proxies for the protocall used in p2p file sharing, is hard to find. Well, annoy ones are hard to find. I'm sure people use stuff to leach out socks proxies, and test at bulk with say, access driver or something, but how many people really know how to do all that ? So, in doing this, you'd be finding people at the source, the torrent itself. Seems like a pretty effictive way to go about it to me.