I think the sub-maximal effort correction part of the algorithm for estimating VO2 max is not well calibrated for me, and I am wondering if it is fixed (the same for everybody) or is it adaptive, so it learns from my runs at different paces how my heart rate varies with pace.
I noticed that my VO2 max goes down slightly every time I do a run slower than threshold pace and up when I do a threshold run.
If I do a workout with a warm up, the performance condition (which is a deviation from baseline VO2 max) starts out as strongly negative, but if I do the same workout with no warm up it starts of strongly positive (which means that performance condition serves no useful purpose for me).
As an experiment, for the last few months I have been doing all my slow runs on trail or treadmill (where they don't count for VO2 max estimation) and all my interval and threshold workouts etc. on track or on fairly level grass (there is a slight hill). My Garmin VO2 max has gone up from 45 to 48, with no other change in my training (I am doing the same 70:30 split of easy:hard runs). Over that time, I have also done a Cooper test each month (on track) to estimate my VO2 max, and the scores have been consistently in the high 49.X or low 50.X, which suggests my actual VO2 max has not significantly increased.
Is this because my heart rate - pace relationship is unusual, so the Garmin VO2 max estimate is biased low?
[in case it is useful, resting HR 37 bpm, max HR 175 bpm, 188 cm tall, 83 kg, and I am in my mid 50s and have been running for just under 3 years]