People are absolutely certain in their own false memories all of the time. So the witness’s certainty is completely irrelevant. https://www.verywell.com/what-is-the-misinformation-effect-2795353
I wonder sometimes if having a more mathematical background would help judges.
Pro: Understanding the idea of averging “certainty” according to a weighting function (i.e. memory before police suggestion should be weighted more than memory after) would help with such reliability judgments.
Con: Despite constantly discussing terms that seem to be refering to probability (e.g. probable cause, reasonable doubt), lawyers, judges, and indeed legislators are never filtered by numerical literacy (at a level relevant to these issues), so honestly it’ll be a long time before it becomes remotely plausible to have that kind of conversation in court.