Updated FTP Estimate outside guided testing

According to Garmins website (What Is the Functional Threshold Power Feature on My Garmin Device? | Garmin Customer Support), FTP Estimate outside guided testing is generated if :

  • 95% of the average power for a 20-minute interval within the ride must be greater than the current FTP setting. AND

  • The average power for that interval of time must be greater than the previous 20-minute power record.

    This would mean that estimated/auto-detected FTP can only suggest an increase in FTP and not a decrease. However, I and many others on the Garmin forum have gotten an estimated FTP that is lower that our current FTP. In my case, neither of these criteria were present when Garmin suggested my new lower FTP. Many replies say that estimated FTP is not based on the above criteria, and instead on some other Firstbeat algorithm. I have not found any source for this statement. So, can estimated FTP decrease or is that a bug? Is Garmin estimated FTP based on the above criteria or some other Firstbeat algorithm? 
  • The user manual for the Fenix 6 states the following:

    "The device uses your user profile information from the initial setup and your VO2 max. estimate to estimate your FTP. The device will automatically detect your FTP during rides at a steady, high intensity with heart rate and power. "

    https://www8.garmin.com/manuals/webhelp/fenix6-6ssport/EN-US/GUID-B40A5128-C382-4F10-AD38-39649C42EF35.html

    I frequently get lower FTP estimates if I do anaerobic or VO2 Max cycling workouts. When I do steady or steady ramp workouts (for example. increase by 5 w every minute), the FTP estimate is very accurate.

    While the 95% of 20mn average power is the common way to establishing FTP through a field test, it was maybe implemented this way on different or older watches.

    I didn't find any documentation on how Garmin/Firstbeat derives FTP on Fenix6/7/Epix2, but it is certainly not based on the power-duration curve (although this is the definition of FTP, BTW).

    My assumption is that HRV is being used to evaluate the second ventilatory treshold (VT2), most likely by estimating the respiration rate and its variation in rate of change with output (Muscle Oxygen Training: HRV with dynamic exercise and Ventilatory thresholds)

    VT2 (respiration) coincides with the second lactate threshold LT2 (muscles/blood), although with no direct causality, and with the "FTP" values of HR and power for convenience.

    I am not aware of other approaches. We start seeing apps using detrended fluctuation analysis (DFA) of HRV R-R alone for VT1 (often called the aerobic threshold), but this approach is not as strongly correlated for VT2 (often called the anaerobic threshold : Muscle Oxygen Training: DFA a1 and the HRVT2 - VT2/LT2)

  • Thank you for your thorough answer. I have the same experience, my estimated FTP increased after 20/60 min TTs and it has now decreased during my training block with 4x8 minute intervals twice weekly. Interesting that hear that the ramp test appear to be accurate in finding a "valid" FTP.