Over all, I don’t really notice that much of a difference between the current update and the last one, but I have some questions that I haven’t exactly seen asked elsewhere. I really don’t want this to sound like a complaint; I’m definitely faster this year than last, and that’s all I could really ask for.
My FTP update is due tomorrow, I changed my remaining plan to different days to accommodate a change in work schedules. Simply changing the plan the day before dropped my prediction from 249 to 244, and I don’t understand how changing the workouts that would occur after my prediction lower the prediction? I understand that potentially changing the plan could force a re-evaluation, but then that brings up my first question where I get a “massive” change for work that has yet to be done. I am potentially missing an endurance ride, but is that one ride the difference of 5 W?
Additionally, I push my TR workouts to my Garmin head unit to control my trainer’s erg mode so I can use Training Peaks Virtual, but do TR workouts for the gains. A week out from this upcoming FTP update, which again occurs tomorrow now, I switched the workouts from Indoor to Outdoor, and before I even did any work, my FTP prediction dropped from 259 to 256. Additional drops I understand to be from not quite meeting the workout goals, but I’m confused by the automatic drop just from marking them as potential outdoor workouts, even though the work isn’t done yet, and will ultimately be done on an Erg mode trainer.
Wild guess: maybe your new schedule is less optimal from a fatigue standpoint and the AI could conclude that a value of 244 would set you better up for the following weeks.
Maybe the AI models have learned that the compliance of outdoor workouts is worse than indoor workouts, and therefore the prediction changes (the AI can’t know that you do the workouts set to outdoor in ERG mode indoor).
Wild guess: maybe your new schedule is less optimal from a fatigue standpoint and the AI could conclude that a value of 244 would set you better up for the following weeks.
It was the night before the FTP update. The schedule couldn’t have been less optimal if it was literally the night before because there wasn’t anything to change, even the following workout was the same, just at a slightly lower power level, and still after a day of rest.
Maybe the AI models have learned that the compliance of outdoor workouts is worse than indoor workouts, and therefore the prediction changes (the AI can’t know that you do the workouts set to outdoor in ERG mode indoor).
Could be. I was under the impression that the AI models were based off completed work though, but somehow I get the feeling it isn’t completely that way since it valued indoor workouts to be twice as efficient as outdoor workouts. Still seems weird to me considering how many people live in states like Florida, Georgia, Missouri, Kansas, where a little bit of route planning, and you don’t even have to ride a highway overpass.