This appears to be a limitation of the way it works by scraping the html pages directly; they don't include the information.

Ideally, to get the complete as-scrobbled information, it should be pulling from the API, but that needs an API key, etc.

Possibly another option is to create a transfer application where a user can do the API calls with their own API key?

Of course, once an audioscrobbler-compatible API is implemented, you'll be able to use existing transfer tools like http://encukou.github.io/lastscrape-gui/