An example of a web-based host cache is Gnufrog. This site appears to update frequently enough where stale information should not be around for long. There probably are other caches and their update times vary. Also, someone could have started downloading the particular file when it existed, was cut off, and is trying to resume the download but the file now isn't there.