Not upset - but want to understand the “why” behind a ~2% FTP drop after completing a prescribed sweet spot workout. So I can change behaviors if needed next time. FTP was predicting 254 watts and dropped to 250 after the workout today 8/13. Feel free to look at my account.
I’ve done antelope +2 and other antelope variations numerous times and this was probably the one with the highest power. Point is there is good comparison data available on this one.
I did add 30 mins z2 at end of workout but I’ve been doing that for almost 2 years no issue.
Nobody can tell you. And it is meaningless. Is the dip real? Was the prediction real to begin with? They are all fake numbers that some algo generated, but can’t explain why.
All that sounds harsh, but is generally true.
My prediction - Eddie will look behind the scenes and tell you your training looks good and ask how the workouts are feeling. Then will tell you to keep up the good work and you should see a nice bump soon.
Yeah I’m interested to see the TR teams review as this is the first time in 6 mos I felt this was erratic. My RPE “hard” was in line with what the workout predicted too. I got up early to get this one in due to a work meeting - I would have loved if I knew endurance would have been a better benefit this day.
I experience similar issues. I don’t have a problem with a prediction that’s less than hoped for, but I do have a problem with predictions at the beginning of the AI FTP update block that seem unrealistic, and drop as I complete workout after workout well. It suggests that the predicted FTP is not playing a role one would naturally expect a “predicted” FTP to play, i.e., get you to that value by following the plan. So, if not a good prediction, then what’s the point behind a predicted value in the first place.
Yes, you can just try each and see what happens to the prediction. But either way, the prediction at the begining of the block ought resemble the actual value at the end of the block. I’m not sure I see the value in RPE-ing the predicted value to some manipulated level that you want/think is about right. That’s what I need the extra intelligence to help with!
Ok, here’s a better answer. For the last 30 minutes of your workout your yellow line (actual power) doesn’t follow the blue line (cooldown power). You extended, added on an endurance interval and the AI is marking you down for adding intensity that wasn’t in the plan impacting your recovery.
Yes but I do that on nearly every 90 min ride, and it’s never dropped for that reason. Further - it originally had me doing 1:15 endurance on Friday as well. I since tried deleting that to see if it helped but it didn’t.
There is your answer . You stepped over the so called “ intelligence “ line one too many times .
Here is some food for thought. At the last day ,end of your plan, you might be 2 watts lower. The next day you might be 2 watts higher. Depends on how much fatigue is in those legs . This changes day to day. Not sure how “AI” can predict this with 100% accuracy .
The Kalshi prediction market favors a drop in FTP if you do extra work. Bet on it.
None of this day-to-day +/- 2W variation is predictable. Far too many human factors. And given that powermeters are +/- 1%, at best, it’s not even measurable. And powermeters are often worse than that because of temperature, calibration, powermeter to powermeter variation, etc.
I’m not convinced that this FTP noise should be front and center in the user experience.
Good point. I guess it made me concerned I’m moving in the wrong direction and don’t know why. With only so many four week blocks before a key event, you want to make the most of them I suppose.
Fair enough, but the workouts are driven by that “FTP” number, so if it goes down below your starting point, you won’t actually have progression in the work in your next block.
I asked Gemini and it thinks TR AI senses micro shifts in cadence and HR. I did a lower body lift two days prior and I could still feel a little fatigue in my legs so I’m not surprised my cadence had a bit more variability in the last sets- maybe minor 3-5 rpm swing. That’s pretty interesting if it picked that up!
