This isn’t a complaint but a curiosity. I’m coming off a week+ of being sick and a recovery week in which I’ve only done 2 endurance level workouts. I had a VO2 max workout scheduled for Tuesday but I changed it to Truchas +1 (essentially 2x22 min sweet spot intervals) to ease back into my workouts. The AI predicted difficulty shows that I have a 92.8% (combined) chance of the workout being moderate or hard. I have a 2.2% chance of failing it. Because of the 2.2% chance of failing it, the program recommends against doing the workout. I’m really curious about such a low percentage prompting a negative recommendation. Why?
First question: how does the workout you’ve chosen compare to recent sweetspot workouts in terms of power and interval length?
Simply put, our AI has very high standards 
There are lots of workouts available for it to choose from, and while 2.2% doesn’t sound very likely, it’s high enough that the AI prefers to recommend other options instead.
My most recent sweet spot workout was a 6.3 level (Eichorn +2) which was 2x20 at 94% of FTP which I marked in the survey as moderate. Truchas +1 is 2x22 at 92-94% and is a level 6.2 although they’re very comparable. I’m not really interested in the difference in the workouts though. I’m more curious about why such a small percentage (2.2%) triggers a recommendation to not do a workout.
Then Eddie has your answer 
