As you know from reading all those other posts, incompletes depend on the source(s) having an exact match. Kazaa used to use a "close enough" match, but that matching system was exploited.

--since LW uses a keyword search to find files, there are many files which "look" to be the same, but aren't. So, those files will start downloading from zero. Old incompletes won't resume unless the source has the exact match.

--the code has changed re incompletes, so maybe some of those old incompletes cannot be resumed, especially if they are quite old. They may be polluting your incompletes. If so, biting the bullet and trashing all those hard-earned incompletes may be the best bet, sadly.

--if the last chunk of a file arrives before all the intermediate pieces are in, the file will look to Explorer as a full file. I think downloads from many sources try to get the early pieces in the right order to help previews, but sources don't always send at the same speed. There is also the need to get the rarer pieces early to help share the load. So Explorer may think the file is the right lenght, but LW knows some of the middle pieces are still coming in.


btw: the chances of completeing an incomplete file from a host that offered a file 190 days ago are lousy if that host was unique --they go offline or change the files they are sharing. The 200 day setting is handy to prevent automatic clean up of incompletes, but then you'll need to manually clean up the incomplete folder.