[riwsky]

“40% replicability” of a field doesn’t distinguish between “4 100% replicable studies + 6 0% replicable studies” and “10 40% replicable studies”, full-stops notwithstanding.