AI FTP Detection Is Now Available for Everyone! 🎉🎉🎉🎉🎉

Good to know and thank you.

I’ve been doing the Kolie Moore progression test but might just see how I fare with the plan following the AI hike. I don’t really care what my FTP number is, as long as my training is at the right level.

Pre AI FTP I was using the KM test or self selecting, as I really underperformed on the ramp test. Like you, I’m not really bothered once it’s working for the training, which I find hard to see how it doesn’t with AT (and honest survey answers!).

1 Like

Going from base to build I did an AI FTP detection 2,5 weeks ago. The result was 1W off where I was standing so I did not accept any changes.
I have recently been sick, and used the Illnessannotation in the calendar. After extending the ilness for more and more days the AI suddenly promted a Ramp test/AI FTP detection. That seemed reasonable, seeing that I haven been doing any training for the last 10 days.
When trying to do a detection (web) I get * Your Recent AI FTP Detection is Still Available* and been told that I can use it again in 10 days and some.
The 28 day window is ok for normal circumstances but wouldnt this case be an obvious exception from the 28 days rule? @SarahLaverty

It is by no means important for me to get the “correct” FTP right now. I was planning to ease in with som easy wo’s and then do some more strenous ones before new test and Base 2 period in 11 days.
When the system triggers a test I would assume it should be allowed

1 Like

After some time off the bike, we will schedule you a Ramp test, to ensure your training is appropriately difficult.

As such, your FTP detected a couple of weeks ago is valid and will ensure that your training continues to progress appropriately. If you accept this FTP detection result, we will replace the Ramp test with another Workout. If you feel like you’d rather an easier Workout to ease back into things, I’d suggest you use Workout Alternates to find a different Workout. Check out this article: How to use Workout Alternates.

Let us know if you have any questions about this!

Not sure about that. Didn’t Nate say something about AI FTP detection being able to down adjust during the 28 day lock in case of illness/injury? So detecting downward would be possible, just upward would be prevented.

Pretty solid, almost 4 weeks of TR workouts, and my FTP went down a watt. With that, my 6 week daily ave TSS didn’t change much during that period, and these 4 weeks were largely on the trainer, or with with HR only on endurance rides outdoors (PM is in the shop).

I would have most likely down the intensity outdoors at a lower HR and RPE, but I would have stuck to the targets, so my post workout surveys might have read easier.

I thought maybe 5 or 10 watts increase going in, but the more I think about it, that might just tip me over to incompletes in workouts on the trainer. Only my first month of TR planned stuff, so I’ll “let go and let god” and see where we go.

Quick question and apologies if answered elsewhere (couldn’t find on a search)

Does AI FTP detection take account of a workout I have completed earlier on the same day?

I’m due to ramp test / AI Test on Wednesday.

I just completed a workout (on Monday). Will it take todays workout into account if I use AI FTP detection now? Or should wait until Tuesday to hit the button or will it account

Can’t seem to find the answer elsewhere. I am doing a long distance tri plan. I have added some extra running into the plan as well so it is 3-4 Rides, 3 swims and 5 runs. In regards to AI FTP detection is it going to take into account the swims and runs? or does it totally ignore them. In theory my cardio should get a bump from those other endurance sports that will help my FTP? Maybe it’s worth still doing the ramp test in my case?

  • No, I don’t think any non-cycling work is considered (totally ignored as you mention).
  • You can test anytime you’d like. But if the AIFTPD value seems reasonable, I’d suggest doing a workout or two with that value to see how it feels. If it’s notably off for your feel & guess, then a test might make sense.

  • Essentially, run the AIFTPD to see what you get and go from there.

  • Maybe, maybe not. Considering the differences between the sports, and the related demands on our muscles and cardio system may vary a bit between them, I don’t know if there is an easy or clear way to connect A+B=C as you suggest.

  • It’s a possible factor, but I expect it’s far more nuanced and difficult to nail down. Just look at the relative difficulty of making this work with a single sport and input and you begin to see there is no “easy” answer here.

1 Like

Apologies if this has already been covered but I couldn’t see anything. If you complete a non TR ride (I know, awful thing to say), does that effort feed into and potentially increase your career level scores?

The only workouts that affect Progression Levels on your TR career at the moment at “TR workouts” as completed on their app or via the “TR Outside Workout” function.

All other “non-TR / unstructured” workouts are ignored from PL consideration. That is set to change whenever we finally get Workout Levels Version 2, that will consider ALL cycling workouts.

However, ALL cycling workouts imported to TR ARE considered for AI FTP Detection.

3 Likes

Thanks for the quick reply Chad, much appreciated :+1: I thought as much but just wanted to check.

1 Like

@Tiiimmmaaayyyy
Except if you associate those with a TR workout. In that case you’ll gain the PL of the associated TR workout. Obvious to say that effort / structure should match as closely as possible.

1 Like

Yup, I tend to leave out the “hack” since it is misused frequently and leads to undesired results.

For those that want to dive into that, they can dig around and find the related discussions.

2 Likes

Although not mentioned in this thread, the associated blog post mentions:
FTP Testing is a Thing of the Past: Introducing AI FTP Detection - TrainerRoad Blog.

Instead of VO2 max being a strict “120%” of FTP, Adaptive Training allows your VO2max and all other zones to be personalized. It learns not only an athlete’s power curve, but also their repeatability in their specific zones.

Can the TR team explain how repeatability has been programmed into Adaptive Training or into Progression Levels? Because you could have high “power at VO2/MAP” etc (without getting into the debate of what VO2 Max power is) and/or repeatability, but there is only one set of VO2 progression levels and they appear at first glance to progress by intensity, length of interval and number of intervals. Or some combination of all three, but not clearly distinguishing between improvements in power curve and repeatability. Which means it is hard to know what workout progressions would be served by Adaptive Training to, for example, Athlete A with high power but poor repeatability vs Athlete B with lower power but excellent repeatability.

Is this just me being over-pedantic or is TR jumping the gun with the marketing vs the current reality of AT using Progression Levels?

I ask because differentiation between highest power and repeatability at Supra-threshold levels strikes me as a very useful thing. Presumably also for Anaerobic and Sprint.

Also, my comments should be in a thread on AT, not AI FTP, but I couldn’t quickly find an appropriate thread and the blog is quoted here. Mea culpa.

2 Likes

You bring up a good point.

I think what you’re saying is you might be really good at repeating 1 min Vo2 max intervals but really bad at sustained VO2 max (compared to everyone else).

Our plans keep you in a specific workout profile so this shouldn’t be an issue. But if you jump back and forth there’s a possibility of what you’re saying could happen.

We also haven’t seen a big jump in completion rates when someone jumps between workout profiles. So although it could be happening, there’s not enough signal in there for us to worry.

Our plans keep you in a specific workout profile

I.e. one plan will have you doing VO2 long repeats at 105-108%, another shorter punchier, another intermittent 30/30-60/60s or Billats, etc… and you typically won’t jump between the different kind of VO2 workouts within a plan block.

Understood.

1 Like

While I get that it’s not going to change that significantly, the 28 day lock was a pain in the ass this morning, as coming out of a recovery week and trying to adjust for next weeks work/ life schedule I can’t update FTP to bring forward the workouts for the long weekend.

Only I had itchy fingers before the 28 day lock, I’d be locked out for 2 workouts at the start of the new block due to having to rearrange. I get not wanting people to update every week, but it would be useful if it was 25 days rather than 28, to give a bit of flexibility at the start of training blocks which are generally the 4 weeks, please @IvyAudrain @Nate_Pearson

2 Likes

Yeah, agree. It does not even have to be shorter period but please make it 1 day precision, not just now() + 28d x 24h x 60min x 60sec. Otherwise it keeps shifting further and further unless I click on detect button immediately once blockout is lifted :slight_smile:

2 Likes

I’ve been using TR for about 3 mo. I hate FTP Tests and the AI FTP seems to great to be true. I see my improvement daily but between each block, the AI FTP seems to increase my FTP my a fair amount… can I really trust it? I’m really doubting myself here so…