Originally Posted by Rowan

I think this graphs illustrates perfectly what the proponents have been saying. That low cadence one is the real clincher for me, but it's also notable that for all the other cadences, there is uptorque evident through large chunks of the phase between 180 and 360.

Sure, the amount of uptorque is small compared with what happens on the downstroke, and may even disappear, but it does exist, based on those graphs, and exists in varying amounts depending on the cadence. Which goes to the heart of people's claims of using uptorque in conditions that seem logical -- hard climbing, for example.

Track cyclists use retention systems because of the very high cadences they require in sprints. Keeping feet on the pedals at those speeds become problematic. But I imagine they use those retention systems to very great effect when starting -- as evidenced by the number of pull-outs that have occurred in competition (on the upstroke).