I’d counter that if we were doing a good job gathering data that these structural biases could be compensated for with more conservative initial numbers.

At some point a lack of decision to take compensating action becomes faking the numbers.