I eventually gave up trying to follow Garmin DSW and purchased an 80/20 HM running plan and a premium subscription to TrainingPeaks.
Initially, during the autumn, my problem was that the watch overestimated my Lactate Threshold and Pace by a wide margin. I did a lab-test in October to get accurate data and set the LT HR to the lab value. But the pace seemed to be stuck on higher values and could not be changed. I was not able to complete LT runs at the DSW recommended pace while the lab tested pace was OK.
Eventually this resulted in overtraining and injuries. I backed off and started doing only Easy runs at around 80 - 85% of the lactate threshold. The trainings come back as close to zero training effect and "No benefit". Not very helpful.
I feel I can no longer trust any suggestions from the watch nor the training feedback. What I get back from TP in terms of running stress and fitness at least seems consistent with how I feel.
I bought my FR965 after 20 years with Polar. Initially I was very impressed. Today I am very disappointed.
I should maybe add that I have a 20 year background as a consultant in embedded SW development processes. I am very familiar with agile development. But what we seem to lack here is making better use of the beta tester feedback (I am a beta tester). And to postpone a quarterly release if the SW is not up to acceptable standards.