I think you might be extrapolating the feedback you got in your previous thread a bit too much. Doing an unstructured ride, mostly Z1/2, but with some short all out efforts sprinkled in is something else than doing a structured ride and adding on Z2 afterwards.
1 Like
Look at the actual riding it’s giving you under each approach and see if you can work out what’s different. Just because you think it’s giving you harder workouts doesn’t mean it actually has …
… and if you don’t want 6 days, change it to what you do want
1 Like
This is a common observation - for many people the balanced plan causes a higher FTP prediction - myself included.
Seems like adding the endurance to the plan upfront helps your situation though?
1 Like
I have an event in five weeks. I have been doing this amount of training for 3 years, there’s nothing out of the ordinary. Every session is marked down. Maybe TR has changed how they rate workouts? The language used to ask, “How do you rate the structured portion of this workout?” But I could be wrong on this. Did an AI search and this came up:
In community discussions and podcast episodes (such as the Ask a Cycling Coach podcast), TrainerRoad representatives often clarify that if a long endurance ride has a very easy warm-up and cool-down, those shouldn’t “dilute” your rating of the main intervals. If the intervals were “Hard,” but the total ride felt “Moderate” because of the long easy sections, you should still lean toward “Hard” to ensure your Progression Levels update accurately for that training zone.
My FTP is 281, last week I did my vo2 structured workout and followed it with 22 min of SS @ 275w. It felt “moderate”. Haven’t done a proper FTP test in years. Historically, though, I’ve found TR to be accurate. Also, TR has not prescribed anything over 12 min for the entire program. So it makes me a tad nervous for my upcoming race, but i’m going to probably add some longer SS efforts to get a feel.
Agreed on weight lifting, I usually have about 20-30 min of actual lifting, 4-6 exercises, 3 sets, 6-10, 8/10 effort.
I thought these might add to the discussion. While balanced and demanding both might yield similar results in FTP prediction, for me anyway, demanding looks to give me more “headroom” where’s it’s not putting the brakes on. My sense from past months is that having more yellow and red on the calendar will lower your prediction, even if it doesn’t actually adapt any workouts.
3 Likes
Fascinating!
Just an anecdote but I regularly add endurance onto my scheduled workouts. AI scheduled 7 hours over the last weeks but I did 10, 11, and 12 hours respectively. Even with the extra work, yellow, and red days, my FTP is set to increase 10.3% tomorrow. Here is an example of a 1:30 workout that was scheduled for last Saturday that I rated as hard. I am curious how it’s going to handle outdoor MTB rides when spring actually gets to the east coast.
I wonder if it’s an outdoor vs indoor thing. My ftp is schedule to go down after a month of 12-15 hour weeks.
It seems like the majority of TR users are doing 4-6 hours a week so when you are more than doubling the volume, the AI gets confused. Once it gets warmer, I regularly do 15-18 hours outdoors. It may not be the best for raising my FTP but I don’t really care. I just like riding my MTB.
2 Likes
Unrelated to OP but how are you adding endurance here? Just manually adjusting the intensity of the cooldown and hitting ‘extend cooldown’ for the time you want?
Exactly.
Meh, I’m now convinced that AI FTP prediction is just a RNG. 
To stress test it, I started a 4 week build phase and stuck, and I mean STUCK, to the recommended workouts to the letter. Rated workouts honestly (endurance were easy, VO2 were hard). AI predicted a 9W FTP increase at the beginning of the block. That prediction decreased about 2-3W each week. Now, with 2 workouts left, the prediction is a 0W FTP change. Glad I stuck to the program 
4 Likes
Might be worth trying for anyone stuck who does not wish to just blindly trust me bro of our ai overlords.
4 Likes
Interesting article. Thanks for sharing.
1 Like
What is probably happening here is the workout level algorithm is way off when it comes to the time factor. It gives way too much credit for longer rides and not enough for shorter ones.
So if you’re doing a TR workout outside and it becomes significantly longer, it gets run through the WL algorithm and gives you too much PL credit.
The AI doesn’t care about the workout level. It’s just looking at power and durations. To see for yourself how the WL algorithm is off with respect to time just look at a future intensity workout and check the “longer” and “shorter” tabs. They all have similar failure rates, but the longer ones have higher WLs and the shorter ones have lower PLs.
The AI should be assigning you progressive power and durations, even if the upcoming PLs are lower than what some longer, past rides are.
1 Like
It’s doing neither.
1 Like
My money is that you actually are doing too much, despite having done it for 3 years, and could make greater gains by following whatever it recommends. (Based on the first couple of months, I may have been overdoing it for 30 years!)
I really would like to see how this works out though….its possible something else is going on. I kinda hope so because personally its been hell not smashing lately
Joe
2 Likes
@AussieRider I’m just getting to this one. I’m going to take a look at your recent training. Do you mind if I share any insights I find?
I’m a bit cautious here. In the past, when TrainerRoad has shared insights, it’s felt a little cherry-picked to fit the narrative. That said, if you’re able to look at my training more holistically, I’m open to it.
Since starting this thread, I’ve shifted my TR plan from 2 days a week to 6 to try to work with the system.
I’m also in a pretty fortunate position with very little in the way of outside stress. No full-time work, no family or major commitments, so I can train when I want and genuinely rest the rest of the day.
One other thing to keep in mind. TR tends to frame TSS on a Mon–Sun cycle. So if I’ve had a recovery stretch with no intensity from Sunday through Friday, then do a group ride on Saturday, which is a full week after my last hard effort, that weekend TSS ends up skewing the data.
Lastly, on volume. Over the past three years, I’ve experimented quite a bit, including pulling volume down significantly, and I didn’t see any upside. If anything, performance dipped.
Thanks. 
2 Likes