On: Intractability of Financial Derivatives

[khafra]

I think when your model consistently gives results indiscernable from random noise in its strongest application (cf. blog.okcupid.com), you can safely conclude that it's of no more value than a model selected at random from all possible models of similar complexity.