IMO, you don’t have enough data in the 90 day window, especially between 20-60min. I’m also guessing your short duration efforts are not real max efforts.
I’d go with 325w from the Ramp Test since it’s early season, and fill in the power curve when it aligns with your training schedule. DM me if you want to dig deeper and I can point you towards a few other things.
Need better data at 20-60m. There was a new chart pushed last week that helps you find the best duration to test at.
Your curve looks a lot like mine at the moment to be honest. I get PBs at 4-10 minutes from a Ramp test, but with trainerroad there’s very little near maximal work, as its all about repeatability or subthreshold. Planning a Zwift session up the Alpe to address my lack of 30+ minute data.
Funnily enough my mFTP is almost exactly the same as my last ramp test. But these are both at least 10-15W higher than I could hold for 30-40m.
This TrainingPeaks webinar does a great job of explaining much of this. Relevant discussion on how to select test durations to improve your power duration model, starts at 38min:
I just used it this morning and it worked well. The charts told me I needed a 34–35m effort so I loaded Reynolds (40m TT) and treated it like a 35m TT. Had a solid effort which pushed my mFTP from 270 to 278, so now it’s only 2 watts lower than Sunday’s Ramp Test result of 280!
Looking back I probably rode it too conservatively as I had enough in the tank to really gas it the last 5 minutes. Definitely was great pacing practice! I used to do 20+ min threshold efforts a lot more before TR and all the sweetspot, so I kind of felt out of touch with the correct RPE…but now I know for the future!
Big fan of that method of FTP testing, it is what I’ve been most confident in. I think he is completely right that throughout the season, anaerobic gains can make it seem like you’re seeing much higher aerobic gains than you think.