> all while two photos are up on the projector, of two children, one with "less than ideal results" and one with "perfect recovery" Seems like you still could

> all while two photos are up on the projector, of two children, one with "less than ideal results" and one with "perfect recovery" Seems like you still could argue the bad outcome have better a priori estimated outcome with that small sample size though? I fear "the other extreme" would be doctors chasing fads.