The sample is biased and intentionally so, the prior probabilities are pulled from general population statistics. Consider the likelihood of an event that should only occur 1/10K times happening 5 times in a row due to random chance, it is incredibly small.