Big question about FTP

So your FTP from all kind of riding was pretty similar and after introducing ramp test it was higher, till the moment you tried to do 20 min test?

1 Like

Sorry your statements are not clear. All I have are:

  • 20-minute test
  • 248W
  • you were wrong

So we still don’t understand, can you help by answering:

  • what was average power for 20-minutes?
  • did you use an app to record? And if yes, what did it suggest as updated FTP?


I do not remember what the average power was during a 20 minute test, at the end of the test a ftp of 248W was assigned. (this is not an average over 20 minutes, this is a recalculation according to the automatic formula)
And with the ramp test the next day, ftp 277W.

yes, you are right, but I have already run the ramp test three times, and the power is always much higher than at 20 minutes.

So basically ramp test seems to overestimate your FTP - and by a lot. If you cannot hold it, even for 10 minutes, this is definitely not your FTP. If you want to use the ramp test try 0.7 multiplier of final minute, not 0.75 and check how does this feel.


Take the lower number - or something in between -

For example ramp tests put me at 245W as I have no V02max or sprint training at ALL. But have been doing slow and steady miles for years and a 20min test puts me at around 270. I can happily repeat 245W for an hour without having to need to be fully rested - thus assuming my FTP is closer to 270. However, I do my v02max training (which i am new to) at 245W.

Its important to be honest and ask yourself - Can I actually hold ‘this’ power for close to an hour (when fully rested)? It should be pretty clear. I’m not sure about holding 270 in my case - but i’m pretty confident holding 260+. 270 MAYBE on a very good day (which is what threshold power is based on).

Also - Through the ramp test is great, i feel a 20 min test gives you a clue to what you are actually capable of under real world conditions.

After almost 3 years of TR and numerous ramp tests with too high FTP estimates, I can confidently say you are far better off training with a FTP that is a little too low than an FTP that is even 1% too high. Error on the side of caution as the latter will end badly for your training and motivation.

After about 2 years of ramp tests I’ve learned I need to stop prematurely and NOT go to failure as instructed, which is about 15-30 seconds before ultimate failure. I’ve learned through trail and error to notice/feel this inflection point, which has resulted in an accurate FTP for me that allows me to complete challenging but doable workouts.

Clearly the ramp test is not giving you and accurate FTP. So chose another method to determine your FTP.

To add a counter point…I bit ago I did a ramp test and got 334W and then a couple weeks later only managed to hold 330 for 18 min. My FTP must be a low percentage of my VO2 (or at least at the time) so the ramp test way over estimates my FTP. I had been really struggling to hold power during intervals and that seemed to explain it. I just fell outside of that bell curve that the ramp test works for.

And again I ask, where does this mistaken notion arise?

FTP is something that you can pretty much dial-up any time you want to. Yeah, if you’re tired from recent training it may feel harder and/or you may not be able to sustain it as long, but you don’t have to taper like you’re going for a world record to “bring it”.


^ This.

No wonder so many people struggle.

Thanks, ok with these two pieces of data:


You learned the following:

  • ramp test appears to be over-estimating your FTP
  • by using the ramp test estimate you didn’t properly pace the 20-minute test
  • if your FTP is actually 250 then starting at 280 means you were starting 12% higher than your FTP (too hard for 20 minutes)
  • if your FTP is actually 260 then starting at 280 means you were starting 8% higher than your FTP (too hard minutes)

Personally I’d try another 20-minute (or longer test), and start conservatively around 245-250W for the first 5-10 minutes and then add power based on feeling.


Yes exactly. Try the 35+ minute baseline test in the link I posted above. It’s very instructive in helping you feel what threshold feels like. I also find it mentally easier than pacing a 20 minute test.

Here’s the link again:


I’m no scientist - TR podcast + forum + dylan johnson is as far as my head goes in terms of self coaching but I have some questions! :slight_smile:

  1. I always thought you should be well-rested before the test - hence the 1 week recovery week after a training block.
  2. The reason you do a shorter test rather than an actual 1-hour effort is both due to it being difficult both psychologically and in terms of pacing too.
    2.1 - If it is psychologically difficult to endure an actually full FTP test on a GOOD day - How do mean that FTP is something you should be able to replicate during a training week given you collect more and more fatigue over the duration of your training weeks.
  3. By ‘very good day’ i mean after a rest/recovery week - which is when FTP is tested anyway? Isn’t this intentional?

This is the first time I’m hearing you should be able to replicate 100% FTP for 40+ minutes on a regular basis… please enlighten me.

It makes sense to schedule after a rest / deload week because that’s when adaptations are made from your previous training block.

It’s no different than with weight training. You might have a 4 week training block where you use your 1 rep max (rm) or 10 RPE use that on your lifting %’s. On your deload week your body heals and gets stronger. When you start again you’ll likely move your 1rm or 10 RPE number up because you’ll be stronger.

That said, I wouldn’t taper and fuel like I’m trying to set a personal world record. I’m certainly not going to be doing that on week 3 of a training block for my third workout that week. It just seems ripe for overshooting your ramp test.

Looking at some data on what all off this is built off of, the correlations between 20 min test and the 1hr aren’t as tight as I thought, but not surprising. My world (geology / hydrology / environmental science) is a lot easier to control than humans, so your test results (if successful) are usually more statistically significant.

With the ramp test, it seems to me the ramp rate can greatly influence the outcome. With racing engines on a water brake type dynometer, the ramp rate (600rpm/s, 300rpm/s, etc) will change the torque (and subsequently the power which is calculated from torque and rpm) values. If ramp test results are % change steps, then I’d argue that the ramp test likely inflates people with lower FTP values.

I think you answered your own question?

1 Like

Again, a myth. Adaptations are occurring all the time; you don’t have to rest/recover for them to take place.

Of course, your performance might (or might not, depending on how you train) still be in the tank until you do, but that’s a different thing entirely.

This is the only study I’ve seen to compare the two.

It reports an R^2 of 0.92 between 20 minute and 60 minute power (obviously multiplying the former by a constant - e.g., 0.95 - would change the absolute value estimated for FTP, but not the strength of the association). The remaining 8% of interindividual variability is probably as much due to differences between subjects in the ability to work above FTP as it is to random “human-ness”.

Right, we’re always making adaptations. Poor choice of word.

As for that study, the sample size is still only 8. I’d argue TR, TP, and Sufferfest have the best data sets, but obviously it’ll never be peer reviewed because they’ll have to make their data open sourced.

The auto world is very good about releasing white papers, either through the SAE or themselves (Honda has a research library that’s free). I would like to see some of that happen, but I’m not holding my breath.

Maybe. Then again, given the aversion of many to longer maximal efforts, maybe not.


I really wonder what % of users have done a ramp test, 20 minute test, and 1hr test around the same time. The longer efforts might have to be pulled from outdoor race sessions (like a TT). My guess, not many, it it would have to be from the same power source. I read that TR can’t tell if a smart trainer is using power match or not, so that further reduces it.

Just simply ride your bike for 40+ minutes and everything will be clear and not estimated. I do not get why someone would willingly do 20 min test (hard as hell) vs longer test - uncomfortable but not different than any hard workout.

1 Like