I've found wget and pretty much every other mirroring software pretty much useless in capturing the whole website. Perhaps I'm using it wrongly, but I had the same problems with windows programs as well.
But what I really wanted was for the tool to check what pages the website actually physically uses, rather than what is on the host file system (which probably has all the same garbage which has been uploaded from my local hard disk), because I'd like to cleanup both my hard disk as well as my hosts files.