[asphaltgambler]

I'll chime in. I think for simplicity sake, I see it as this. Take an engine that makes 100 HP @ the crank. It is coupled to a conventional rear wheel drive setup sans vehicle for this discussion. If measured at the drive wheels the loss might be10% / 10 HP.

Now take the same engine which makes a verified 200 HP @ the crank, bolt to to the same existing drive train connected to the wheel dyno. Assuming the trans/ diff would tolerate a smooth application of twice the power - I don't believe, as a matter of fact some one would have to prove this in real world testing, we'd see the same % of loss.

Now in actual on the road testing, the same drive train rated for 100 HP would most likely fail at some point with shock loading, torque. I'm sure the friction would increase but I don't see how it would double prior to failure.