It also might help to have two day inbetween the anerobic and Vo2 days. The AI seem to like it when our hard days are spaced out.
2 Likes
That is a good point, and I have considered that, however saturday + monday for my hard efforts just fits really well into my schedule.
1 Like
So what is the marketing purpose for those that it reduced the FTP by at the start? Do you think everything in life is a conspiracy theory?
6 Likes
It is nothing like ChatGPT. It is a machine learning model not a LLM.
3 Likes
Lots of good info here. One thing I would add is to change the way you think about AiFTP. Don’t look at it as a measure of your fitness. Look at is as metric that the TR system uses to give you the best workouts. Like others have said, I’m sure you gained fitness over these four weeks. But the model still thinks this is the best baseline to give you optimal workouts based on how you’ve been performing. This is exactly what a good coach would do—see that you had some tough workouts and suggest not raising your FTP or target power too quickly. Focusing on a 4 week change in your AiFTP will get you wrapped up like a pretzel (especially if you are watching the prediction after every workout).
Not to be nitpicking …you completed the workouts but a few weren’t perfect which is why it scaled you back. That is not a criticism, just a reframing. I see a lot of people saying something similar or not following a plan and wondering what happened
Good luck!
8 Likes
Thank you very much for your input, and you are probably right; I should forcus less on the FTP.
I totally agree with focusing less on FTP. Remember that the TR AI doesn’t know FTP, it just looks at power, HR and RPE.
I turned AI FTP Prediction off just because I found myself fixating on it too much, especially when I had to miss workouts due to illness.
2 Likes
The simple answer is you went from training five days a week to three.
Big picture… try to schedule weight with the least impact but also consider short term challenges (impact on workouts) for long term gain. If progression is gradual your body will adapt and be stronger.
Perhaps I should try that
I guess you are right - I’ll hang in there, and hopefully see improvements on the trail in the spring/summer
1 Like
The consistency in your plan is 
.
Going from some of the what I ran into with my training.
With the volume that you are doing would you be able to added one more Hard day per week and then add in an endurance ride on the weekend if you have time and can keep that consistency. Even if it’s in phase say one hard day now then after a month added in that endurance ride.
Either way you’re stronger now than at the beginning of the plan and by your AIFTP holding at the same level, your time in zone for the workouts will be getting longer with is good in its own right.
1 Like
Thank you very much for your input.
I will try to stay consistent, and then I will hopefully see progress.
I don’t know if it’s been mentioned, but you seem to be in specialty phase. FTPs didn’t use to get updates during that phase in order for the algo to optimize your current fitness for the upcoming event. It’s also not structured to increase FTP to begin with, but rather to help you express it, hence the emphasis on Vo2 and anaerobic work.
On another note, I saw someone comment about your anaerobic workout dropping from lvl 4 to 2.5 : that could well be because lvl 4 is where you start with in the anaerobic zone (if that was your first such workout) and you rated that very hard, so the algo maybe thought it should ramp you up from a slightly easier start point. It’s not a cause for concern in my opinion, just the algo getting to know you. I’m a bit more curious about the 1.8 afterwards.
2 Likes
What does the pdc look like pre/post plan?
Thanx for the reply.
The new AI prediction, continously updates the predicted FTP; that started out being a +11 improvement, and then it declined week by week to 0. So while you are absolutely right that I didn’t aim for the best improvement in FTP but rather general fitness, it is still surprising to me, that it decreased the prediction week after week, and ends out at 0. But yes the algorithm probably needs more recent data ‘to get to know me’
I just finished the plan, so I don’t really have any post-data to compare with yet - but you are right I should probably look for more granular data using the entire power curve.
Just look at it from before you did the plan compared to what it is now, that will show you any improvement even if it’s not FTP.
1 Like
I just had a look, comparing dec 2025 + start jan 2026 with end jan 2026 + feb 2026 - there is no improvement. But all my training in the last period are TR workouts, so I guess that is not the best way to compare?
I mean I would expect some sort of pushing out on your pdc somewhere 
. If there are no personal bests at any point along the pdc then that could indicate sustaining your current level rather than increasing it but as always it depends.
1 Like
Yeah, that is a little strange. Hopefully I will see improvements after the next 4 week block.
1 Like