> This seems like a really straightforward thing to test. If I get a thousand people together, have them each bite 10 human skin analogues, identify the partici

> This seems like a really straightforward thing to test. If I get a thousand people together, have them each bite 10 human skin analogues, identify the participants and each of their bite records with a GUID, can bite mark analysts map bite records to biters and with what accuracy? Not quite the same thing. Because that assumes you already know it was one of those people. More appropriately you should ask the labs "does this bite mark fit any of these 100 people", where none of those people actually caused the bite mark. Will the lab recognize it was none of them? If they just falsely identify one out of the hundred, that extrapolates to millions of people.