Comparing actual FTP Test to the AI FTP Detection Feature

And in TR there are the PLs to adjust for different strengths of athletes, shouldn’t that take care of “*** strong athletes” for the given FTP?

at the expense of overzealously training other zones? if your “FTP” is 20-30W over actual, everything in z2-z4 is going to be far too hard… imagine your threshold PL is like 3.0 and you’re doing 8m intervals because they’re actually suprathreshold.

it doesn’t produce the intended adaption, and making it “easy for threshold” doesn’t change that.

better off using the real FTP and either ignoring power suggestions for vo2+ or bumping your PL up high for anerobic riders

1 Like

only if you buy into the idea of using FTP to set targets for work above FTP. But that comes at the expense of over-estimating targets for threshold, TR style over-unders (between 95-105%), and sub-threshold work like sweet spot.

Despite some flexibility introduced with AT (high WL for vo2max are above 120%), it sure looks like TR is still clinging to the idea its ok to use FTP to anchor all efforts.

This is a solved problem outside the TR world - use 20-60+ minute field tested FTP, and define individualized targets for above threshold.

2 Likes

After a couple years of foundering between coaches and TR, I am convinced that in my case, I need a full 60 minute test to accurately test for FTP. My best 20 minute test was over a year ago and it put me at 294 watts. 19 minutes of that test had me absolutely pegged at around 300 watts, the last minute I somehow found another 100 watts.

If I was honest with myself at that time, there is no way I would have been able to hold 294 watts for half an hour let alone a full hour. I wasn’t sandbagging that first 19 minutes either, it was 10/10 for me. But as with most workouts when I see that 01:00 turn to 00:59, I somehow find another gear and send it.

That 294 watt FTP sent me into a year long spiral of stagnating and eventually burning out. My coach at the time said I was completing all my workouts but he didn’t really have any suggestions as to why I felt like crap all the time.

2 Likes
  • That is a clear red flag that the test is flawed.
  • Those two prior statements don’t mesh. You might think you were pegged, but if you truly manifested another 100w for anything more than a few seconds, there was “gas in the tank” mentally, physically or some combo. All that to say, that nailing a test like that is anything but simple or guaranteed.
  • If that is the value you pulled from the 20m test mentioned, I would consider that highly suspect. I’d have had a tendency to throw out the last minute and do the 0.95 multiplier on just the 19m AP. At best, round up a percent or two considering you had a kick in the tank.
  • If that 294w came from the full test, I think that was a mistake. But that is admittedly a guess based upon minimal info and no actual review of the test in mentioned.
  • How have these longer tests gone for you? Specifically, are you nailing them with “flat” average power and minimal peaking that you describe in the 20m test above?

  • How long have you been training with FTP derived from the longer tests?

5 Likes

By default Frank Overton uses 90% for anaerobic athletes, and 95% for diesels. Assuming you used 95%, then using Frank’s 90% you would have estimated at 278W which is aligned with your earlier comment of 270ish FTP.

1 Like
  • Those two prior statements don’t mesh. You might think you were pegged, but if you truly manifested another 100w for anything more than a few seconds, there was “gas in the tank” mentally, physically or some combo. All that to say, that nailing a test like that is anything but simple or guaranteed.

I looked at the full test, the last minute I averaged 363 watts and peaked at 422 at the end. My heart rate from 10 minutes to 19 minutes averaged 175, my max HR is 180 and anything above 170 I am usually on the floor gasping for air. I honestly feel like I was going my absolute hardest for those 19 minutes and the last minute/50 seconds was pure adrenaline.

  • If that is the value you pulled from the 20m test mentioned, I would consider that highly suspect. I’d have had a tendency to throw out the last minute and do the 0.95 multiplier on just the 19m AP. At best, round up a percent or two considering you had a kick in the tank.
  • If that 294w came from the full test, I think that was a mistake. But that is admittedly a guess based upon minimal info and no actual review of the test in mentioned.

My coach at the time used the standard 0.95*20 minute power. In retrospect this was definitely not accurate.

  • How have these longer tests gone for you? Specifically, are you nailing them with “flat” average power and minimal peaking that you describe in the 20m test above?
  • How long have you been training with FTP derived from the longer tests?

I haven’t done a full hour test but I have gone 50-minutes at what my coach considered SST at the time, between 270 and 288. I failed the workout once but I completed it a few days later. I had no joy about completing it, the workout absolutely destroyed me. The following week, my coach gave me 60-minutes at SST and I could not complete it. Looking at my calendar around that time is where I started going downhill and never really recovered.

  • All of that, coupled with the flawed use of the full test show a group of mistakes coaching IMO.

  • I suspect you see this now, but failing to complete even a few of those “SST” workouts is not typical or expected. They are signs of an incorrect FTP or incorrect progression of power targets and/or duration. In any event, they should drive some decent review and conversations to find out what is off, because something surely was.

  • Sounds like you have better perspective now, so that is good to have as takeaways. Always be learning :smiley:

1 Like

Coaching mistakes, and going back to the original post whatever you want to call this:

Sounds to me like 88-90% of best 20-min power is a better estimate of TRusername’s FTP. And that TR Ramp overestimates, as does AI FTP detection.

agreed, always be learning!

Which reminds me that I’d love to hear any feedback you get from TR related to this use case. I am hoping they can leverage this and other cases that seem to be off like yours, and improve the detection setup for others with similar use and/or training profiles.

1 Like

I did hear back from the TR support team, they were very prompt about it. The support member said my AI FTP was indeed inflated largely because AI doesn’t work well with outdoor workouts, of which the majority of my workouts are.

1 Like

OK, that makes sense in the current state. Thanks for the update :smiley:

It will be interesting to see your AIFTPD results once they release Levels 2.0 update, since that is supposed to do a better job of analyzing and assessing PL’s and such for all the outside stuff.

1 Like

I think one key learning is that short tests - ramp or 20-min/8-min tests or CP - are estimates of your ability to sustain a longer effort. Before understanding some of Coggan’s thought process that led to the term FTP, I figured “why not simply do a longer effort?” and figure out how long I can go at my best attempts to pace such a long effort. And that was pretty interesting experiment back in 2017. However just like there is no single % for everyone to make TR ramp (or 20-minute test) produce an accurate FTP estimate, picking 60 minutes is not automagically going to make the FTP estimate more accurate.

Appreciate you sharing all the twists and turns of coming up with targets for your training!

1 Like

woah. 309W for the 20 minutes, but 363W on the last minute? I guess another learning is either you don’t know how to pace 20 minute tests, or its the wrong test for you!

1 Like

Kolie Moore and Kyle Helson just put out a pretty informative podcast on FTP testing shortcuts that is worth a listen. Podcast link is here

1 Like

I have a paced 20-minute FTP tests poorly in the past so I know how they feel. This last test did not feel like a poorly paced test. Frankly if someone can explain how this test went from a physiological standpoint I’d really appreciate it, because I have no idea how that worked.

1 Like

Yeah I dunno and don’t feel like speculating. All my untrained eyes see is that you are likely better off doing a ramp or 8-min or 20-min overpaced effort to come up with pacing for a long test. My coach has me do things like a 10-min hard ‘what you think your 20-min pace should be’ pacing effort ahead of a 20-min field test. And for a longer test that never happened, a few months ago I did a 32-min pacing effort. My fitness has NOT changed much since then, based on some 6-min threshold efforts on Monday, so I’d still use info learned on 32-min to pace a longer TT / field test effort.

Max efforts (ie tests) require preperation, skill, experience and motivation to perform well relative to your physical capability.

If FTP estimates from actual tests are based on the assumption that you perform the test well, then you’d expect the estimated FTP to be a little higher for most people.

I did my 1st AI ftp last week and it put me up 2.5% from 219 to 224. I saw it coming, and it was pretty much in line with expectations. Actually the highest % increase I’ve had, but I think my last ftp ramp test was not as good as it could have been.

The subsequent sessions at the higher ftp are noticeably that bit harder, but not that much that I cant get them done.

Happy.

1 Like