This all came back to my mind this week as it was testing week for me
Yesterday the dreaded 5 mins and today the 20.
Honestly, nothing beats the feeling of climbing off and seeing a new PB. Feeling like my work paid off, that I actually did it and proved I could hit new numbers, and the motivation I felt to get stuck into the next block, was incredible. 100x better than getting a pop-up saying I should increase my FTP setting or an email from TP. I already knew my FTP was up as I could feel it in my riding but nothing beats that post-test euphoria for me despite the fact I dread them the days before
I also think this whole ‘its a waste of a workout’ is total BS as well. No way a max 5 or 20 is not training, especially when you can have a 5 min recovery and hit some z2 or z3 for 20-30 mins afterwards. If we’re talking about a KM TTE test then even more so. I dread them but I’ll keep doing them.
The interesting part will be if we are a signifigant amount off the prediction and how you handle it. Will there be a survey to give feedback to take care of things that don’t really mean your adapting different then expected:
different equipment (sure, your gear is accurate for good power numbers, but switching to something new may read different)
change bike fit
feel tired (need more recovery)
If you are adapting at a different rate than expected should your progression levels be changing at a different rate then it was to challenge you more or less?
We know there will be a lot of forum members who will not like their detected FTP. And there are so many different definitions of FTP that we’re going to get flack.
But…what we will look at is lots and lots of data to see if detected FTP + AT = Productive workout. If we’re doing this correctly, we should get appropriate RPE responses post-detection and get people to the right levels.
Even if we’re off by a few watts, with AT that’s still OK. I believe that we’ll be more consistent in our predictions than people are at executing a ramp test (actually going all out, being rested, lots of motivation, etc).
But I fully expect a 1,000 post thread fueled by 200 athletes saying how it doesn’t work for them. And that’s OK. They can still test other ways and we’re not removing any options.
Another way to put it: If we get 80% of athletes choosing AI FTP Detection over the ramp test AND their post workouts have the correct RPE and failure rates, then that would be a HUGE win. Once it goes into production I’ll most likely post some metrics in this thread (no promises though).
I think we’ll hit higher than 80% though.
AND, there’s another project going on right now to further increase the accuracy of AI FTP Detection and AI FTP Prediction.
I’m in the 80%+ and very excited for this development. I do not test well so I have been estimating for myself but it is hard to separate ego, optimism, self-preservation, and reality.
Huge salute to the TR team for their vision, belief, and determination
This actually makes me more excited about this feature. If I’ve done ANY recent VO2 work I over-test in the ramp test to the point where (pre-AT) I used to start failing workouts and burning out. It took a long time to work out how best to address it. With AT it is better but I still think it’s not ideal - it’s still demotivating to start struggling and have your workouts adapt downwards, especially when “adapt downwards” is often in terms of interval length, etc, rather than the actual power target.
If the FTP estimation accounts for this and gives a more achievable result where the workouts actually feel “right”, it should be celebrated, despite the ego hit.
Just like any thread on an internet forum that even lightly mentions FTP will at some point devolve into arguments about what FTP actually is. It’s basically a law that can’t be avoided at this point
I need this now I did a 20 minute max effort today, and TrainingPeaks & Intervals.icu both said I increased my FTP, and their estimates are 1 watt off from each other. So perfect time to check TR’s estimate
While I get this to some extent, I do think there’s a difference between repeatable intervals that might have you feel like your downing at the end vs. working to the point of exhaustion where you actually have to stop pedalling. Maybe it’s only a few % more, but it is much more mentally draining. I can do it, and I do it regularly, but if I can avoid it, great. Especially when it doesn’t always give me a good training FTP estimate anyway.
Over time I’ve come to think that if I’ve done recent VO2 work it gives me higher, otherwise it’s fairly reasonable. I think my ability to push at the very end of the ramp can ebb and flow throughout the year in ways that aren’t accurately reflective of what my threshold is.
No…no we need to lay low at the start, stealthy. Then when everyone is sufficiently harumphed into a lather…that’s when you come in with some sweet sweet 2007 memes.
And for the coup d’état maybe you can post the dancing baby from the 90’s. The farting frog would be legendary too, but we don’t want to crash TR servers.
@Nate_Pearson
Curious if your data on FTP estimator has been benchmarked against the Ramp Test.
Are you using the Ramp Test as the accuracy target as you tweak the estimator algorithms?
Or are you saying the estimator may be better than other test methods. If so, how could “better” be quantified I guess.
I think this is going to be a great option!
But for the first couple, I’d still do a Ramp test to “judge” the accuracy. If it’s close, in the future I’d just count on the estimator. If it’s off by a lot, then I’ll have to see what gets me more productive workouts.
I’m interested In seeing how the TR FTP estimator compares to the WKO5 and Intervals.icu estimators, now that I’ve just done an all out 20 minute effort. I’m not using 95% of this effort to set FTP, but to use it to feed the WKO5 model. I’m using the WKO5 mFTP to set sub-threshold training zones, which is why I’m very interested in seeing what the TR estimator spits out.
How does the WKO5 model work? I know the Intervals.icu one well - it basically just looks at best efforts along the power curve and estimates FTP from that:
But I don’t know the WKO5 model.
Sounds like TR’s model is predictive (i.e., taking into account work outside of max efforts and predicting on that basis, using some kind of ML model) rather than analytical (i.e., analyse for best effort and apply a multiplier).
Agree it will be super interesting to see how they compare, I wish I could actually see the data and not just what Nate chooses to share! (I know that’s not practical…).
I set my FTP based on feel more often than I test (as I have found over time that the numbers are basically identical anyway), so it will be interesting to see how this compares.
I am sure your predictions will be good for the vast majority. Was more focusing on how with ML when the model doesn’t predict the real results there can be something worth looking into. (when the numbers are signifigantly off, not a rounding error off) Not because someone complains, but more the automated backend just looking at what they actually did vs the prediction of what they should have done. Then when you have a data set of those that predicted too high and those that predicted too low you can look into general trends of what is special about them. Some of those data points should be thrown out because the change is based on things you don’t know about (bike fit change, different equipment, etc) and things you might be able to predict but still easy to ask (they feel tired or worn out) so could be very useful to have a survey for ramp tests.
Assuming always good manual RPE responses from surveys seems like a bad assumption as it is almost the same as assuming someone complaining their predicted FTP is right. I’m not saying to go off of them contacting support or complaining in the forums, just based on the data you see.
My prediction is that those who are newer to training and/or starting at a lower power will be harder to predict as their rate of increase will be greater then those who have been training awhile or already high in power. They are further from their bodiy’s potential and have more areas where they can improve (the different types of physiological improvements like mitocondrial growth, capillary beds, etc) but will be hard to know how fast they can improve. Go under in your prediction and they won’t improve as much as they can but if you go over in your prediction you might get closer to overtraining. I’m guessing being new to it also means they won’t be as good at rating their RPE for different workouts
I did a 20 minute max effort today, and TrainingPeaks & Intervals.icu both said I increased my FTP, and their estimates are 1 watt off from each other. So perfect time to check TR’s estimate
