Differential vs. Integral W' bal Formula

[James Bell]

Is there a reason why one would be preferred over the other?

  

[Nathan Townsend]

Yes.  I prefer differential formula at present because tau is dynamic with respect to recovery power whereas tau for the integral model is static based on the entire ride duration.  Therefore, lets say you do a short unders/overs type bout for 10 min inside of a longer 2-3 hr ride at Z2.  The static tau for the integral model will grossly underestimate the actual value that should occur during the unders/overs segement.  

At the moment, a big limitation of both models is that both use first order kinetics of the recovery and depletion.  I think we need to move to something a little more complex because for example, the time course of recovery of central versus peripheral fatigue mechanisms is likely quite different, so a first order system just can't handle the complexity.  I spoke with someone at ECSS in Vienna who said that a computer engineering/maths nerd colleague had trouble fitting Skiba's biexponential model as descreibed in his 2012 paper. So it's not gonna be an easy task methinks.