So Jukebox, do you agree with the studies presended in the link - at least for a conventional gearbox with fixed ratios?
http://www.stanford.edu/~voloshin/lhowwhy.html
Would you then agree that the power peak would also be more important than torque in my example with my hypothetical lossless 20 speed gearbox, even though the revs change very, very little between one gear ratio and the next?
If you answered yes to the questions above, then why do you think it is torque that matters more than power, for a CVT with cone and pulley or whatever?
All gearboxes have powerloss (CVT too I'm sure), but I can't see how that matters to the maths of power and torque. Hopefully if there is something I'm missing here, you'll let me know, but the way I see it, at any given instant, the CVT has a fixed ratio too and the normal laws of physics apply to the CVT example.