Here is what he wroteTorque analyses that involve net torque (SpinScan, SRM) must make a certain assumption to do their analysis. The assumption they are forced to make is that NOTHING is done on the backstroke between 6 and 12. Having made this assumption they can then say what they see is what the rider is doing on the downstroke (and of course, "everyone" knows that is where all the power is generated so that is all that counts). The problem is, of course, that having made the assumption that NOTHING is done on the upstroke (between 6 and 12) that it is not possible to show that the rider is starting the power stroke at 11 since they assume he is not. Hence, I think it is quite clear that "Providing the data on the net torque from both cranks is" NOT "more than sufficient for this purpose." In fact, it is demonstrably useless for this purpose.