I've seen a number of match/id errors go away if I remove the extra text after year. I access my files directly and include extra info in the filename.

For instance:

'Twas The Night Before Christmas (1974) Rankin Bass
Chances Are (1989) {} Robert Downey Jr. Mary Stuart Masterson Cybill Shepard
South Park S04E01 The Tooth Fairy Tats 2000.avi

Randomly, this will interfere in the id of the file. The YEAR in the South Park example triggered a failure to identify. It would be great if we had a token to cutoff the extra stuff. Prior to all name cleaning, you chop from the token to the end off, then start the existing process of resolving the name.

'Twas The Night Before Christmas (1974) !-- Rankin Bass.mp4
Chances Are (1989) !-- {} Robert Downey Jr. Mary Stuart Masterson Cybill Shepard.mkv
South Park S04E01 !--The Tooth Fairy Tats 2000.avi

This would let us have all sorts of info at the filename level without risking interference. I've had to delete quite a few to get a reliable ID.

The XML comment seems unique enough to re-use. "!--"