Good catch -- the link to the code was wrong (it was linking to an old version). I updated it to point to the new NCP version.

That's weird that it does several different types of moves in a row into the same register -- there must be some kind of implicit calculation going on, or maybe it's just a bug.

I'd use "git blame" to bug-shame whoever wrote that heterosexist code, but I don't think they were using safe source code control practices back then.