Question about AI Training behavior during illness, Masters plans, and taper logic

Training
1

Hi all
longtime TR user here, and I’m hoping to sanity-check some experiences and better understand how the new AI Training system is intended to behave in a few edge cases.

I’m not looking to be negative; I’ve gotten a lot of value from TrainerRoad over the years. That said, I’ve repeatedly run into situations where I’m not sure whether the intended behavior of the system matches what’s happening in practice, especially as a Masters athlete with a long training history.

I’ll break this into three related questions:

1. Illness after a recovery week: best way to rearrange without resetting

I recently came off a scheduled recovery week and then got sick. To avoid prolonging the illness, I spent ~9 days doing only Z1/Z2. That meant:

  • I missed 3 planned intensity sessions

  • One week ended up with only 1 intensity workout

  • The following week I missed the first intensity day and only got back to intensity mid-week

So in practice, I ended up with about 1.5 weeks of intensity across a 3-week span, and overall TSS was well below plan.

My question:

  • What is the recommended way to handle this in the new AI system?

  • Is the expectation to reset the plan, or to push the recovery week out, or something else?

  • Right now, the calendar logic still wants to insert a recovery week even though I haven’t accumulated much fatigue yet.

I can self-coach around this, but I’m curious what the AI is supposed to do in this scenario.

2. Masters plans and FTP downward spirals

Last fall I tried a Masters plan. As expected, it removed an intensity day and reduced overall TSS. What surprised me was what happened next:

  • Reduced intensity → reduced TSS

  • AI revised FTP downward

  • Easier workouts → even lower TSS

  • FTP revised downward again

There didn’t seem to be a mechanism to hold FTP steady in the absence of sufficient intensity signals. Once I exited the Masters plan, FTP improved again.

Questions here:

  • How does the new AI system prevent this kind of downward spiral?

  • Is there a concept of “maintenance” FTP when intensity is intentionally reduced?

  • How does the model distinguish intentional load reduction from loss of fitness, especially for Masters athletes?

3. Taper behavior (especially for race-critical events)

This is the one that’s bitten me hardest.

Last season going into Masters Nationals, the plan over-tapered me significantly. I went into the race with a form around +20, which for me is well into detraining territory: legs felt flat, no muscle tension, no snap. Performance suffered.

When I asked about this at the time, the response was essentially to self-coach the taper. That’s fair as tapers are individual, but it does raise the question:

  • How does the new AI system currently handle taper weeks?

  • Is there any differentiation between “fresh” and “too fresh”?

  • Is the model aware that some athletes perform best closer to neutral or slightly negative form?

For context, I later rebuilt my own taper with a more gradual volume reduction and minimal intensity loss, and it worked far better. But there’s no way to express that preference in the system.

TrainerRoad works extremely well for me most of the time, probably 90% of the year. The problem is that the edge cases (illness, Masters-specific adaptations, tapering for key races) are exactly when guidance matters most.

I’d love to understand:

  • What the intended behavior is in these cases

  • What’s currently automated vs what still requires manual intervention

  • Whether the new AI platform addresses these issues differently from previous versions

My calendar should be linked to this profile.

Thanks in advance, genuinely curious to learn how you’re thinking about these scenarios.

2 Likes
2

I’m a masters cyclist too, and I share almost your exact same concern about tapers. I actually perform better at the end of a TR training block than I do after a TR taper. I start to fail what should be very doable workouts when the load comes off.

Last year, I was able to salvage the start of my A event by manually adding volume in the final week to get my form number up to 10, rather than 20.

Unfortunately, I think the taper structure programmatically built in by TR, and the system as a whole is not looking at how we handle tapers individually.

But… if we can get the prediction to fall on race day, then I think we could play around with the schedule in the last couple of weeks to maximize predicted FTP. That way, the AI would be looking at how we respond individually to the final prep.

That makes me wonder @SeanHurley , is there a way that we can shift out FTP detection date so that the four week interval falls on race day?