Alex, good thought, but how would you determine how hard is hard enough? I would assume this would be subjective, but would like the opinions of others.

Random surveys of a statistically significant population could yeild a mean, but would it truely be meaningful?

How hard must it be to be optimal?