I think the source of confusion is that the "response delay divisor" isn't a delay at all. This parameter is the inverse of amplification. If you dampen a signal more then it will oscillate less.