Why don’t they allow bad statistics in libraries?
Because it’s too noisy
(Ramzi & Bethany's noisy data inspired this one.)