- 10:23 pm - Thu, May 8, 2014

Ward-Smith CP vs WKO4 FTP

If you use the correct version of the Ward-Smith model and let the CP tau2 be a fitted parameter it turns out that you get something that gives you an CP essentially the same as the WKO4 model.

P(t) = W’*/t*(1-exp(-t/tau1)) + CP*(1-exp(-t/tau2))

At least it does on Andy’s data set.

I find this a bit interesting as I had this discussion with Andy of which was the appropriate version of Ward-Smith to use for fitting MMPs and he didn’t seem to believe me. He also didn’t seem to believe me that the Ward-Smith even had a W’ equivalent term but it does.

It is not the most consistent equation to use in terms of how easily it converges without getting the start parameters just right, and the above has a couple years missing because I don’t really feel like coding up the brute force algorithm. The issue being you’ve exponential feed backs coming from above and below.

Either way though, if you don’t feel like waiting for WKO4 you can use the Ward-Smith and get all the same actionable information in terms of Pmax, W’, and CP. Or at least it looks that way and at some point perception inevitably becomes reality.

And since, according to Andy everyone has the same fatigue resistance as a percentage of CP so there’s really no need to fit anything out past an hour or so. Just look at CP and call it a day.

Of note, Ward-Smith has been around since 1985 and is I think based off and very similar to Lloyd even before that, go figure.