Maybe I should clarify my thinking here. I mean that if a jury are told that the DNA matches the prosecuted with 99% certainty, then most people would interpret that as the person is guilty without doubt. But if you said that theres a 99% statistical probability of two of the 57 persons in the same court room where born on the same date, then I think most people would be very doubtful, and even if you checked and it showed that indeed two people in this room where actually born on the same day, people would see that a coincidence, not statistical proof.
Assuming DNA matching has 99% certainity. Are you saying if I get DNA samples from 57 different people and match each of the 57 samples in all possible combination of pairs (so 57*56 tests), then there is 99% chance that I'll find 2 samples which will match?
Someone should do this experiment. I think it's very scary if this is true.
Comparing everyone's DNA vs everyone else's would be like the birthday paradox. However in court they've already picked one sample (the evidence)