I've done some checking for other parishes (michael + Ballaugh)- the base error rate for the burials is about 3 to 4% - in some periods this rises to near 10% so you need to be very careful in blindly quoting Brian's index - most errors are just the final digit wrong (often 9's replaced by 0) usually in the year rather than any other field so in general not easily noticeable but the swopping of digits is another tho thankfully much less common error. Of course there are other errors in that Brian merely indexed the transcriptions by the IoMFHS and these also have errors including misreads of earlier (ie pre mid 18th C) entries in parish regs (+ that old favourite of the year change of 1752)