I’ve been spending a lot of time reading the forum with this new release and it’s sort of baffling the negativity some have been showing towards these changes.
TR: “With our new system, your FTP value may be a bit different than you’re used to, but we’ve tested it and our new system gives the best training outcomes for most people”
Critics: “My FTP is different than what I’m used to, this doesn’t make any sense, I’m not willing to try it because the FTP doesn’t match what I”m used to”
Why not just give the new system a try and evaluate it based on the outcome, not your expectations?
I did some more research on my FTP test and I did a critical power test that involves a 3 minute and 12 minute test.
3 min = 296W
12 min = 240W
Which equated to a 221 FTP number.
Some would say apparently that Critical Power is typically 4–6% higher than your 60-minute FTP.
My coach used this for my FTP number which is quite common but others would argue that this should be reduced especially if workouts are too hard and may lead to more fatigue.
The standard conversion is to take 94% of your CP to find your FTP equivalent. Calculated FTP: 221W×0.94= 208W which is near enough to the TrainerRoad detection.
With this new information I will now be accepting the new 208 FTP detection to help avoid burnout and hopefully maintain a steady progression.
Thank you to all the friendly responses received and no I’m not a troll as someone else suggested.
I was someone like many others, confused by FTP tests and calculations and it just goes to show that FTP can actually have a broad range according to the tests and required adaptations.
Just want to say though there is nothing necessarily wrong with the way your coach did it.
Just as I’d always argue you should use the TR number for TR training… it makes just as much sense to use your coaches number for workouts prescribed by them… they will be using the number that they feel works for their workouts.
So an update since lowering my detected FTP to 204 from my originaL coached FTP of 221.
AI recommended my first workout on Sunday and with the lower FTP I was expecting an easy or easier workout. And as some predicted it was about right.
It was a prescribed as a hard session and it was borderline hard/very hard, so I rated it very hard and this week I’ve got the same workout prescribed but AI has lowered the intensity because I rated it very hard.
My learnt takeaway from this, is don’t expect an easier workout with a lower FTP number, in fact it was just as hard as anything I had done previously with my original higher FTP number.
Those of us just getting on with training aren’t making a post unless trying to help when they can.
To counter the negative posts, so far in my experience it’s been detecting easy/hard/v hard pretty spot on and I feel my threshold workouts are nice and productive. Hard but not so hard I’m fucked for the next day. I have said elsewhere I got a drop of 10w, was sad but decided after some reading to roll with it as I’m a computer engineer not a coach so assume they know more than me.
Not quite. It’s exactly the same intensity of workout (3.1). When you performed it this week you exceeded the power targets a few times and the AI recognised this and scored it as a 3.3.
Oh, I’ve got that wrong again, I thought AI was being clever and adjusted it for this week due to my rating but thanks for the info, everyday is a school day especially with TrainerRoad.
If you’d rated it Hard it probably would’ve given you a different, slightly harder, workout this week to stretch you. As it is, it’s giving you the same workout in the expectation (I’m supposing) that you’ll rate it Hard next time and then it’ll progress the next workout.
If you go back and change your rating to Hard, you can see that in action as it adapts your subsequent workouts. You can always change it back to Very Hard afterwards.
I’m quite often in two minds how to rate a workout, so I do both ratings while looking a week ahead and sometimes what I see makes my decision for me.
Just tried adjusting my rating to hard and it increased the endurance workout intensity very slightly and kept the remainder of the week exactly the same.
However when I re-entered my very hard rating it didn’t revert the easy workout back but instead changed the original hard workout to an easier one. Small differences but not making much sense!
It seems to me that these days many people recognise that whilst the notion of an FTP can be useful it is not really a 100% reliable universal metric that can encapsulate all you need to know to compare the ability of two riders or even fully characterise how your own ability might be changing over time. For instance many can accept that one person might be be able to hold FTP for 20min whilst another may be able to hold it for 40min; or that one person “scores” a higher FTP on this test rather than another test whereas for someone else the results my be opposite; or that one person has a VO2max power at 120% of their FTP whereas for someone else it may be at 110%; or that one person has a higher FTP outdoors rather than indoors and another does not; or that one person can ride for 20min at FTP after 2hrs endurance and another cannot etc.
We have been told that the new AI system does not actually use any “FTP” value as such as a data point to determine your workout intensity or progression and against the above backdrop it’s not too hard to see why that makes sense.
I personally would have been happy for Nate to have gone down the road he has said he really wanted to take and abandon the reference to FTP altogether but I can see why he will have been concerned as to how this would have gone down with the user base (though that now seems to be a bit of a “damned if you do, damned if you don’t” scenario). With the benefit of hindsight TR may have been better served beginning a move away from using “FTP” as an anchor point for its plans when it first introduced AIFTP - something along the lines of “If you don’t like doing FTP tests TR AI will now generate an AI ‘workout baseline number” for you which will be used to determine your workouts and track your progress and which for many people will broadly equate to your FTP (which you are still free to test every so often if you want to)”
If you are following a plan but are unhappy with your new reduced FTP number just try and enjoy the confusion and angst it causes your friends when they can’t seem to drop the person with the paltry FTP - I know I will having just “lost” 33W off mine.
I suppose that a much simpler algorithm would likely just switch between the two same set results as you switch RPE back and forth as you expected but AI is somehow a different ball game (and may possibly have different approaches to dealing with a retrospective increase vs a decrease in RPE or even detect your uncertainty - AI is a complete mystery to me). What you are doing is understandable curiosity as to what impact different RPEs may have as you feel your way around the new system but of course in reality retrospective changes in RPE will not be the norm and certainly retrospective back and forth changes shouldn’t really happen in an ongoing training scenario so perhaps the AI simply hasn’t been “trained” on how to deal with this (whatever that means).
Yep, definitely curiosity, I was borderline to rate the workout hard or very hard, so chose very hard. Helvellyn suggested for playtime to see what happens when the hard was selected instead to see how the algorithms change for the recommended workouts ahead. Toggling between the two showed different workouts each time but the result is probably the same with some minor tweaks.
