My guess is that the true underlying treatment effect of Provenge in the 9902b patient pool is about 1.25. Even in a PSB world that is below the raw (no Cox Correction) HR for 9902a (the *second* trial).