Originally Posted by
globecanvas
Also, wow, the training peaks NP and TSS are way different from what I calculate for this ride, by over 25%. FTP is set the same.
I know there are two different commonly used windowing methods that are used for NP, one of which is used by Strava, the other by TP and CA. I know they are most different for rides with a lot of variability, like this one. I guess this is why some people always have ATL numbers that seem crazy high to me.
This doesn't apply directly to this, but is perhaps illustrative of some of the issues. I know if I drag my srm files directly into WKO, vs using device agent I get different numbers.