Agree. Because I think it’s making people miss what is working better and 80% of the posts are about it.
This is where @Nate_Pearson could return and give some actual data. His videos and sharing of data was awesome during the beta. I assume they know what the general factors are that are causing very optimistic predictions. I’d think it looks better to admit there is an issue and these are the variables that influence it. Then just have hundreds of these posts a week.
Mine is pretty optimistic this month, but I can see why that is and I’m pretty confident it will drop. But if I do hit the workouts on the schedule then the value makes sense. I have a custom training approach with Sweet Spot set as aggressive and threshold as demanding but endurance at moderate. As I like sweet spot and plan on doing my endurance outside with my son once my hand is cleared to ride outside, so they will be chill.
Exactly the same thing happening over here. Crazy high prediction that lowers, mostly in the last week or so. I’ve come to not expect the gain it’s predicting and assume (hope) I will improve a couple of watts over the 4 week block. So while I’m not holding out for that increase and facing huge disappointment, I do wish it was at least (more) accurate.
I appreciate the AI is beta, and I won’t pay too much attention to minor fluctuations, but this has intrigued me:
Yesterday’s prediction for may 3rd was 256w, a workout was in my calendar ready to do
I missed the workout due to work commitments
This morning the prediction for may 3rd had dropped to 255w. Makes sense as I missed a workout.
I added a workout to today (was a day off, but I wanted to make up for yesterday). I used AI and set the type/duration to match the one I missed yesterday). That dropped my AI FTP to 254w (I think because I’d be training 3 days in a row)
I deleted the additional workout, and my predicted FTP is now 257w (so 1w higher than before I missed a workout!
So, missing a workout, adding one then removing it has bumped up my prediction by 1w. woohoo! I like this kind of “training”
Very similar experience today. Am near the end of a 4 week block, next FTP detection will be next Wednesday. AI FTP prediction has been swinging around between a 16 to 25W increase for the last few weeks, and before this morning’s workout was showing as a 19W increase. This morning’s workout is the last interval session of the block (next week is a recovery week so everything between now and the AIFTP detection is rest, easy or endurance day). Workout was a VO2 session (Bolton Brown -2), did it outside and absolutely nailed the intervals - all the VO2 intervals were completed in middle to top end or just a couple of watts above the 20W range that the Garmin gives me, HR was normal, felt pretty good and was actually having to hold back a bit to hit target power on some intervals as felt I could have gone quite a bit harder. Only bit I didn’t execute well was that after the VO2 intervals finished there was a 17 minute Z2 interval and due to route planning my watts were below Z2 for a lot of that as it was descending or going through busy traffic areas to get home. Marked the workout as Hard (some of the intervals required some focus) figured the AIFTP prediction wouldn’t change much if at all, but no it dropped straight from a 19W increase to a 6W increase.
I don’t really disagree with the AI prediction - always looked optimistic and a 6W bump is still decent. But if it’s going to drop 13W after a single well executed workout and no other changes to my plan that I can see in the next 5 days (and this isn’t the first time it’s happened for me or clearly for others) then do have to question what the point of even having it is.
I’m pinging the team on this one again. I’ve just done some more digging and have new info that I’ve shared with them. I’ll follow up as soon as I can!
