I'm raining for the Grand Canyon R2R2R. Curious about the productivity of my training, I plotted it.
As a runner, I train primarily by power using my Stryd foot pod. As such, Stryd will score each run with a running stress score. I can then calculate the weighted and unweighted rolling 7-day and 42-day averages. From there, I can get the stress balance.
Stryd defines their productivity training boundaries different than TrainingPeaks. Here are both side-by-side.
I don't know which boundaries are "correct", but as an observation, Stryd's are more uniform than TrainingPeaks.
It seems TrainingPeaks is more forgiving on what counts as productive training. But TrainingPeaks is also stricter on "race readiness", giving you a tighter window to make sure you're not losing fitness.
Both platforms have thousands of users. I'd be interested to heard the arguments from both on why they chose the boundaries they did.