Brian - yes I agree, as the Manx part of IGI is based on the Rolls Office transcription of Parish registers done by Manx who knew how to decipher Manx names they are generally very close to 100% accurate though some entries were apparently missed most are mentioned in various IoM FHS journals - some rescuable from Bishop's transcripts; re non-Manx names the few I've chased seem slightly more error prone but sample is too small for valid statistics.
BUT it is now littered with the junk entries and it takes significantly extra time to use the on-line IGI - I find it easier to use the old microfiche (not the latest version which apparently is as bad as the on-line IGI - I have been promised an article in the next (May) issue of IoMFHS journal by another researcher who also commented to me on the same failing and she speaks with many years experience.