The expression you'd want to use would be:

ftp(meanmax(power),60)

This would create a history chart of mFTP based on a sliding 60 d window (instead of the default 90 d). After that, though, you'd have to manually set your FTP using the Athlete Details page.

The reason that WKO4 doesn't automatically push mFTP onto your sFTP is that although the model is quite accurate, it isn't necessarily perfect. Furthermore, mFTP will wiggle up-and-down a bit as older rides roll out the back and newer rides roll in the front, whereas you really don't want to be micro-managing your sFTP that way.

If you set hard defined points (every 60 days, or 90) would that not increase the fidelity of the PMC chart by adjusting the sFTP? Are the generated variables for TSS factors into calculating mFTP (which would induce some recursion error)?

Definitely a good idea to stay on top of significant changes in your FTP, since it enters into the calculation of TSS, and hence CTL, ATL, and TSB. The way the history function works, though, is that looks back XX days (where XX = 90 unless you specify otherwise), scoops up all of the data between then and today, and builds a mean maximal power curve on those results. This means that it can potentially change on a daily basis, as older personal bests for power roll out the back, and new personal bests for power roll in the front.* Thus, even though there is no recursive error, if you try to respond to every little wiggle by changing sFTP (or if the program automatically set FTP to mFTP), you can still end up chasing noise, rather than true changes.

(*Note that because of the way WKO4 - and also other programs - collates such data, the response to training- and detraining-induced changes is asymmetrical. That is, as soon as you set a new personal best it will be reflected in the model parameters, e.g., mFTP. However, when your fitness is declining there will be a delay before it is reflected in the results.)