If I'm reading this correctly
This is very old news. The tracker protocols (actually just a HTTP/URL protocol for announcing you want to join the peer network) have had the ability to specify the IP address you're connecting from for ages. Not all trackers will support the IP= field, but it seems that most do. It's very handy if you're sending all your tracker requests through a VPN, tor network or web proxy, which seems to be the main reason it's supported in the first place.
Once again, I'm pretty sure the code described in the article is based on old knowledge. There's a good paper describing how (I think) this technique was used to get DMCA takedown notices sent to a network printer:
http://dmca.cs.washington.edu/dmca_hotsec08.pdf
The Reg article is also quite a bit off when it talks about using this for "anonymity". In fact, if you want to connect to the swarm, you have to have a valid IP from which you can connect. Once you're in the swarm, the tracker knows about it and will gladly tell anyone who makes a request to join later. All you achieve by feeding fake IP data to the tracker is to increase the amount of unnecessary traffic that will be generated by peers trying to connect to random machines and making it more difficult (but not impossible) for eavesdroppers to say with certainty whether a given IP is actually part of the swarm.
The next logical step for eavesdroppers is to actually attempt to connect to the sites themselves and see if they're actually accepting incoming bittorrent protocol messages. Of course, the other side can also make things more difficult by picking machines they know will be listening/conversing on a given port, so eavesdroppers might end up with a few more false positives than they'd like. Plus, if there are enough fake IP messages spread out to enough popular active torrents, the effect on the targetted IP address could be tantamount to a DDOS. I'm actually surprised nobody's gone and exploited that one so far.
Although it's not mentioned in the article I linked, there's also the possibility of using the Peer Exchange feature of bittorrent to achieve pretty much the same spoofing behaviour.
Finally, I think I'll go anon on this one cos I'd written my own scripts to do this spoofing since around, oh, Hadopi 1. Not that I've ever used them to bombard French government sites with masses of torrent traffic, but since it seems like such an obvious thing to do, I don't want to get blamed for the actions of any other (hypothetical) script kiddies who've had the same idea. Better safe than sorry.