I kinda know the answer, but want to confirm it. My FTP was lowered in today’s detection from 250>233. For reference i haven’t train indoors since the middle of march (two months now). I did only two short z2 ride in trainer road this week due to bad weather. AI FTP asked me to measure my ftp and i did of curiosity. It was lower. I am certain that it cannot access it from 2 short rides after i haven’t use the app for two months. Am i right, or my fitness got worse. I am crushing my PMs my a large margin outdoors.
Are your outdoor rides from the last two months uploaded on TR? If so, maybe have TR estimate your TSS from HR data (if you used one). Otherwise the lack of data has most likely impacted the AI’s ability to detect properly. If you’re starting to do more inside riding now it might be a good idea for you to do a test or manually adjust the FTP closer to what you think it is then manually adjust as you go
Hey, I don’t think TR uses hr or tss from outdoor rides. My outdoor hr is higher due to heat and humidity. I do not plan to ride indoors until autumn . Then I will manual test it.
It does. You can get TR to estimate TSS from HR data. Search the forum and you’ll see how to do it as I can’t quite remember how to so it ATM…
I’ve done it myself recently and it works
Yes, it calculates tss based on hr, I know that . I don’t see Link between tss and ftp calculation without data. It seems impossible.
So you’re asking TR to assess your FTP without controlled training efforts? I didn’t see whether the OP rode outside to power or not, but I bet if there’s no power data from those outdoor rides then yep I wouldn’t think TR can really truly do much with that - even if it does estimate TSS.
And even if you are riding well, smashing Strava segments or whatever, have you actually done efforts that demonstrate you can hold (for example) 2 x 20 mins at sweet spot or repeated 30sec on/30seconds off? Or other power durations for repeated efforts in the same session? Relatively stochastic efforts in isolation I wouldn’t have thought are enough for TR to assess your FTP like it can from controlled turbo sessions.
If you think the new FTP estimation is wrong then plug in some sweet spot or threshold sessions that are 8+ and see how you find them. TR will quickly catch up if you do that.
If you don’t feed TR with decent data, how can you expect it to do anything other than account for you not having trained consistently for 2 months? Doesn’t mean it’s necessarily correct, but what else can the system do?
For reference, here is the TR support article and excerpt that is relevant:
What AI FTP Detection Needs to Detect Your FTP
In order for AI FTP Detection to have enough data to accurately detect your FTP, you need a minimum of 10 completed TrainerRoad indoor workouts. If you haven’t completed enough workouts, don’t stress—you’ll see how many more you need when you click the button to use the feature. Once you’ve completed these 10 workouts, AI FTP Detection will analyze and incorporate all indoor and outdoor rides with power data, whether or not they’re TrainerRoad workouts.
Key point above is the latter bold section about rides needing power data to have impact. All signs above point to the OP not having power data outside (title & other comments), so the basic take would seem that TR is not seeing any of the recent outside workouts (even with estimated TSS from HR) as related to FTP.
As ever when an estimate is in doubt, a real FTP test may be the best action.
I am not tech guy. Not sure how AI stuff works, I needed confirmation of my observation. It should be wrong as I firstly thought. Things are advancing in AI maybe the stuff I see on internet about AI is not correlated to TR system. I do much more sweet spot outdoor than indoor. [only with HR and Rpe]. My tte has gone through the roof. Incorrectly thought that ai is advanced to measure ftp only from 2 short rides which now sounds stupid.
- If the support info I shared above doesn’t cover your question(s), I suggest an email to support@trainerroad.com for the quickest and most direct answer.
Thank you, I feel kind of embarrassed from myself for asking such irrelevant question … and I will probably delete the thread. .
It’s not irrelevant. Totally legit to ask since it’s not obvious what is “driving the bus” here until you dive into some relatively obscure links You aren’t the first and won’t be the last to wonder about things like this. Good for people to see now and in the future if they search on older topics.
Hey there,
Great answers in this thread already – I just wanted to confirm that it does seem very likely that your recent AI FTP Detection looked low because most of your rides didn’t have power data over those 8-9 weeks.
While AI FTP Detection does take your HR data into account, your AI FTP Detection will be much more accurate if you can use a power meter. If there isn’t much (or any) power data for the algorithm to use, then the AI FTP Detection result you get may not be as accurate as it could be.