4 Successive AI FTP Decreases - Good plan compliance. Help/Roast Me Pls

Looking at your calendar, e.g. on March 20 you did “Pioneer -3”, which shows no sparkles, meaning it was not planned by the AI. There are other workouts like that in your calendar. On a related note, on March 17 you were supposed to do James (Endurance 5.0, 1:30) but went for a 1:11 easy ride instead.

I didn’t even look after I read this part. There’s nothing wrong with doing things this way, but it probably explains a lot.

Can you explain this more? I’m not sure I’m following what you mean

I mean that “Decent consistency and volume” coming after you said you make frequent changes says to me “I didn’t really follow the plan”, so I’ll leave it to the TR folks to comment on what had a positive or negative impact on performance results vs. expectation since we can’t see the details, just the calendar overview.

Welcome :slight_smile:

why are illnesses no longer marked in the TSS/week overview dropdown in the calendar view?

You mean on the TSS chart above your Calendar? May be a design thing.

You mention the sparkles, but those sparkles are on the alternates when chosen through the workout alternates menu on the calendar, so I’m not sure I follow you.

The sparkles on completed workouts mean that you followed the exact workout that was prescribed. Changing it to an Alternate, means you still deviated. Not a bad thing, but it’s not what the plan recommended. However, the good thing about these Alternates is that the AI is choosing the best alternatives for the original workout so you don’t end up doing something too tough or too easy.

I think my main gripe is that I get presented with incredibly optimistic FTP gains at the beginning of every single block, train with decent consistency and volume, and then end up with a drop in fitness after 4 weeks of training.

I wouldn’t get fixated on the FTP number… There are other fitness gains like completing harder workouts, feeling stronger, recovering faster! A number is not your entire identity, plus if you were to do an FTP test every day I guarrantee that number would look very different on a daily basis.

3 Likes

Yep, the TSS Chart. If I scroll back, illnesses are marked with a red dot. But more recent entries on the same display are not.

Thanks for clarifying re sparkles.

It’s all good and well to say to not fixate on FTP, however TrainerRoad is very FTP number centric. What you’ve mentioned doesn’t really narrow down much on why I am seeing such optimistic FTP predictions, training at a moderate to high volume and then getting decreases through successive blocks.

What you’ve mentioned doesn’t really narrow down much on why I am seeing such optimistic FTP predictions, training at a moderate to high volume and then getting decreases through successive blocks.

I think in my explanation it does :see_no_evil_monkey:

When you first get the AI FTP Prediction it’s assuming you’ll do all thats laid out. If things change along the way (like the ones I mentioned) then it’s only natural the prediction will change… It also has more of a butterfly effect at the beginning than at the end of the 28 days.

Have you given permission to share your data on the forum? I’m curious also.

Do they need to do more? Seems pretty clear that permission has been given seen as they have also linked their calendar.

I can’t see the calendar, only their career page and some other things.

This would seem to indicate some inconsistency, but hard to draw with confidence.

In regards to TSS…. What inconsistency are you seeing here? Looks solid to me.

Go to that Career page and replace “career” in the URL with “calendar”.

Surely the numbers speak for themselves, I don’t know what else to add.

What does the word ‘consistency’ mean to you?

What would a “consistent” TSS chart look like to you?

I’m not interested in this game.

Thanks Obda

@TheGearedGuitarist

You’re still showing an increase of over 2% in this cycle. It looks like you’ve added some rides to the planned rides?

The last cycle showed a tiny drop in FTP, it’s hard to see on my phone but here in the week leading up to detection;

You’ve struggled on the vo2 workout due to stress/motivation, and on the threshold workout you haven’t registered an RPE - these are likely indications that your fitness had not improved as much as expected.

What’s the response when you adjust your plan to more or less rides per week?

They explicitly gave TR permission to share their calendar and relevant information here. I don’t think they need to do it again?

Not a game. I am seriously trying to understand what you are seeing so I can learn. As a coach… If I saw a training calendar with this training stress I’d call it consistent. But maybe I am missing something. If you don’t want to elaborate what you are seeing why did you post it?

I have to stick up for myself here a little. My TSS chart is pretty spot on for 3 weeks on, 1 week off of training as prescribed by TR.

in reply to @JoeX , I can’t add any more days (6 days a week at the moment). Taking a day off my training plan results in a predicted drop in FTP. The screen shot you posted of my TSS is from almost a year ago. Not relevant to the current discussion.

in respond to @Caro.Gomez-Villafane - let’s change approach. What would you change in my training to see progress as opposed to going backwards. I think it is not unreasonable to change the occasional workout using workout alternates so that the structure able to be completed outside (I mostly do this with sweetspot workouts), or if I have more time to train in a day (the very occasional Saturday threshold work. What can I be doing differently within the structure of trainerroad to see better outcomes for the time and dedication I put in?

Not a direct answer to your question, and not trying to be snarky. Why don’t you try turning off AIFTP prediction and doing you best to follow the plan. If you make some alternate picks go ahead. Then focus on the progression over three and six months. My $0.02 is people are WAY to focused on the prediction. They should be focused on the 28 day detection, the quality of the workouts delivered, and longer terms gains. Remember AiFTP “Prediction” is in beta.

If I were TR (Nate and Jonathan), I would make AiFTP prediction an “opt-in” feature as a beta feature. It’s off for everyone unless you opt-in with a disclosure that it is beta (and you can’t endlessly complain about it on the forum :wink:).