Let me start this by saying how much I love the new AI training as well as the predicted FTP. It offers that carrot that helps me be focus and stick to the structure rather than banging out my own unstructured rides. I’m significantly fitter and faster since starting in January.
The only flaw I’ve found in the system is that predicted ftp seems to be overly optimistic - “this will be your future ftp if you complete every workout perfectly, mark them all as easy and add some special unicorn sauce”.
For example, this is my last week before the ftp update. It’s a recovery week so all the workouts are very easy and manageable. I did my workout today exactly as planned but for some reason, my predicted ftp dropped by 5 watts.
My suggestion is that you make predicted ftp more conservative. I’d rather be pleasantly surprised on the update day than slightly disappointed .
Will a more conservative model give the useful nudges still?
Take this weekend, I can train my longest ride Saturday or Sunday. I assumed that Sunday was best (based on plan spacing) but the prediction is showing a better outcome training on Saturday.
What would a more conservative model do here? Give neither an advantage (thus I potentially lose out on that useful nudge) - or start to suggest one is a negative?
Appreciate a swing at the end can be frustrating but I see it as useful for finding a general direction and making small marginal improvements to decisions that add up over time.
The sensitivity is helpful for playing around and seeing what changes do.
Unfortunately the estimates are a good demonstration of behavioral economics and loss aversion. People hate loses more than they like gains. Having that prediction feels like something we are owed rather than something we earn.
I think that rather than making the model more conservative they should show a range. I imagine the median value wouldn’t be centered in that range but it still might be more helpful than an individual value. Also could help if you want to optimize for floor, ceiling, or just let it ride.
Yesterday vs today based on that 1 workout that was completed exactly as prescribed.
Don’t get me wrong, I am still super happy and that is a still a big gain, but as one commenter suggested, I would prefer to overshoot my predicted ftp rather than fall short.
What is the spacing? The only reason it “might” make a difference if you need to be recovered for a Monday VO2 max or something. I have certainly seen long endurance rides impact a VO2 max session if its a day later, but two days later is fine.
I agree with your general point, so please don’t take this as negative, but in this specific example, it’s kind of impossible to know if you really did “exactly as expected” because of the nature of a “solo ride”. Who knows what would have happened if it had been a structured ride that it could judge to the nth degree.
Like I said though, I do agree that it needs a band or to be more conservative earlier. Far too many people are being disappointed by large last minute swings after a month of following the plan. There needs to be a way to make it a more positive experience.
Gotta say, I like it. So many things I thought would increase my ftp seem to lower it. That has been the biggest eye opener ever. FTFP baby, don’t try to get fancy (that’s my self talk nowadays)
Guys, here’s the hack to not being disappointed. Have nothing on your calendar so you get a massive drop in FTP. Then every workout and ride you do slowly bumps the prediction up as you progress through the calendar.
How so, it was marked as complete as planned. Even so, even if I had completely skipped the workout entirely it was just an endurance workout in a recovery week. It should not be that impactful right? I would understand if I failed a key workouts.
It has no data to analyze. It wasn’t a structured workout that it could at least look at time in zone with, so even if it takes your word for it that you performed perfectly, there is nothing to analyze since it’s just a “solo ride” marked as completed. Did you do better than it expected? Worse? Exactly on the dot? Was your hr normal? Higher or lower than usual? Did you take any breaks? It has no idea. I wouldn’t expect an accurate analysis from that. Throw in that it tends to swing pretty aggressively near the end of each 4 weeks.
Again, I don’t disagree with you about it being demotivating and maybe needing a band rather than an exact prediction, I just don’t think it’s a surprising drop all things considered in this specific example.
I’m actually in the process of doing this. I’m selling house and moving across country. My trainer is boxed up and I won’t be riding for two weeks, so the prediction plummeted. Now, when I go for a run my predicted FTP is going up.
I would love just 10w extra in real FTP this year