Currently, when I e.g. rip a CD and submit the results to CTDB, then my own submission is already counted as confidence when re-ripping the same CD.
Re-ripping may have several uses, e.g. ripping with different programs (which in practise turn out to find different issues which the other don't find, I've had quite a number of such cases). Or simply using different settings, e.g. in EAC one could rip once with C2-errors once without.
Counting one's own submission is especially bad for very rare CDs where (if at all) there's typically only 1 or even just 2 submissions.
If the one is then my own, than bogus data is written to the logs and a confidence is assumed which isn't there.
AFAIU there is some user ID stored in CTDB submissions, so I think that could be used to filter out one's own submissions.
Or one could simply store the submitted results locally,... perhaps the MB discID + the actual CRCs,... and if these two come up together then don't count the own submission (i.e. subtracting 1).
Sure that isn't perfect... but better then nothing.