It's pretty common to see the 20 min test erroneously predict FTP for about 60% of the people who do it.

And that's among the people who do the 5 min blow out before hand.

There are software programs out there than can reduce the need for formal testing as long as your doing some informal testing through out the training cycle

---

Brian Stover USAT LII
Accelerate3 Coaching
Insta

There were a few exercise physiologists active on this forum who developed and advanced the concepts of FTP and testing approaches. They could elaborate in massive detail. (I am not sure how many are currently active.) My suggestion is track down the people with first-hand expertise and ask them.

why is that not good enough?

This.

My personal 1:1 is that lots of average joes speak in terms of erroneous 20min tests. "My ftp is blah blah blah" from some dodgy 20min test they did on an over-reporting wheel-on trainer without doing the 5min "blow out" before the 20min test.

So to avoid this, I now just adjust my workouts as I go to match my fitness. If I set out to do some work and can't finish it, it's a bit of a clue. Adjust a few % and go from there. For the "bad" workout where your guess is wrong, you can always hit the +/- % button mid workout OR you can extend or shorten the sets. If your guess isn't "that bad" I'm of the opinion that works better than beating you head against the wall trying to hit an erroneous ftp test result in workouts.

My go-to if I "must" is I'll go ahead and take the TT bike for a 40k spin OR do a trip up the Alpe in Zwift. The Alpe in Zwift for me being longer than 45min but less than 60min, so within that "full ftp" duration.

It's a tried and tested method, that is convenient, repeatable, and relatively simple to perform (time and equipment). It gives a very easy to interpret number, which can be tracked and has a pretty good correlation with overall cycling performance. Not only that, but this number can be used to set training zones - sounds obvious, but remember plenty of tests (e.g. vo2 max) don't. You could argue the training zones aren't perfect, but they are a reasonable starting point (fwiw my LT1 has always been measured right around the top of my zone 2 based on coggins FTP based zones).

If you are going to criticise a test you should probably suggest something better. I wouldn't argue against lactate being more accurate, but it has its own downsides in being a practical measure for most. True 1 hour FTP may be better, but how many people are seriously going to do it?

It was never meant to estimate an athletes maximal lactate steady state, that's probably one of the reasons why.

Edit: So Training Peaks describes it as MLSS but that's not exactly what I remember from AC when he was here.

My point is pretty straight forward. I understand a basic FTP can provide an estimate of an athlete’s work capacity at threshold and I utterly believe that is a good thing for most athletes.

However, overestimating power out by (conservatively) 15% for an entire training block based on a more or less accurate FTP test means that training volume and training intensity are going to be, on average, 15% more demanding than needed to drive adaptation. More volume and more intensity over time to me sounds like a turkey voting for Christmas: it is overtraining waiting to happen, and overtraining is, indeed, is a fairly common issue among endurance athletes.

Those who truly care about improving performance and seek marginal gains to improve a personal best should strive for a bit more of a data-driven approach. That’s what the Norwegian are doing. They have no secret ingredient. They use data.

I found better results by incorporating a 5min blow-out at the current FTP before commencing the actual test. It is a pretty common protocol used in research and it does seem to increase accuracy. Anyone out there willing to try it, feel free to share your results and feedback. Would love to hear from you.

---

Antonio Squillante
Ph.D (s) CSCS*D RSCC SENr
Clinical Exercise Research Center (CERC)
University of Southern California

Where is your evidence that 20min FTP test performed correctly overestimates power by at least 15%? As I said before my FTP test lined up pretty perfectly with my lab test results, I doubt I'm a huge outlier in that regard.

Let's say you really do overestimate FTP by over 15%. As soon as you do a structured work out with some intervals it's going to be unmanageable and you would presumably scale it down.

Again, you can criticise FTP, it's easy always easy to scrutinize the norm and it's clearly not perfect. Until you put forward a better approach, it's kind of moot though.

This is worth a read!
https://www.trainingpeaks.com/...w-testing-protocols/

Also tune in to his podcast episode 1 is exactly on this subject.

I think this is only a concern if someone is a robot and executes a set training plan based entirely off of %FTP. If a session is too hard as planned, adjust it. Maybe this is a problem for those that don't know anything about training and just blindly follow Trainerroad/Zwift workouts, but I think most people that are coached or are planning their own training understand the need to adjust when necessary.


Power is data. So is heartrate and RPE. Beyond these data sources, I'm not sure what is available at reasonable cost and extra effort to the typical athlete. What are you suggesting?


Are you proposing a 5 minute interval at FTP in addition to the 5 minute all out interval the is part of the protocol in Training and Racing with a Power Meter before the 20 minute test?

I will agree that Zwift and other software basing their FTP suggestions off of 95% of 20 minute power alone is probably giving a lot of people inflated estimates of the power they could hold for 60 minutes. Again though, it is just a number. People with inflated FTPs will just figure out that they can't really do X by Y minutes at Z%FTP and adjust either the workout or their FTP.

The original protocol includes a 5 minute all out before the 20 minute all out (with a rest in between). What protocol, specifically, are you interested in having people try?

Genuine question - if there is a test, maybe your proposed one, that gives me a number that is exactly my power at MLSS, what do I gain by having that versus whatever number an FTP test spits out that isn't exactly my power at MLSS? How does it change how I train?

1. Doesn't that really depend on who's training philosophy you use? If you are using an approach that limits time at and above FTP, then sure an athlete will have problems but they also will not be able to complete workouts as prescribed. If you are using a sub-threshold approach, the athlete will likely not be over trained, just have the wrong energy systems targeted.

2. That is the FTP test proposed by Hunter Allen and Any Coggan to estimate FTP. The straight up 20 min test is not. Threshold is near MLSS, but it is not the same.

Isn't FTP just a way for us all to speak the same language? So what if the 20 min test over estimates it by a bit, then just prescribe the threshold intervals at 90-95% of FTP instead of 100%. I agree it's over estimated using the 20 min test but I just use it as a common language for us all to speak when talking about the intensity level of the workout or race powers.

This forum is awash with stories of people who did not adjust their #s and ended up fried, burnt out, with subpar performances and for many it's happened 2,3, or even more times.

The problem is your # today is not your number tomorrow or the next day. It can be 295 on the day you test. It's not a static #, that goes for MLSS or FTP.





FTP is just 3 letters/words people use to describe a point where if you're working above that point the duration you can sustain work/power is going to be drastically shorter than if you're below that point. Which, more or less, describes MLSS. FTP isn't a physiological point like MLSS

When scientists and non scientists discuss FTP, MLSS and/or CP it often seems like people are talking as if in different ballparks. Everyone is in the same ballpark. Some are looking at it from the 15yd line and some are looking at it from the bleachers and then describing what they see.

The reality is none of these numbers, MLSS or FTP is a fixed #. Who cares if you have an exact FTP # or a MLSS #?
They will vary a bit based on stress, sleep, caffeine, fatigue, training or lack of.

It's better to be 10w under that # than 5w above that number. You and jwott made great points about setting it below the number and being willing to adjust based on how things are going. Some athletes though are hey I got 268w and if I'm at 270w I'm above it and working vo2 and if I'm at 259w I'm not improving my threshold.

---

Brian Stover USAT LII
Accelerate3 Coaching
Insta

If you're not a coach or a researcher and you're having conversations with enough people about their FTP and how they calculate it to determine what lots of people are doing, you need to rethink things, IMO. Why would you talk about FTP with anyone outside of maybe your training partners?

How are you assessing the accuracy of your FTP tests, such that you can conclude that the 5 min blowout increases the accuracy? Perhaps the benchmarks you're using to validate your FTP tests should become your FTP test.

Not to mention the number of people even here who definitely have wacky powermeters.

Some reulgar poster not long ago mentioned he estimated that like 20% of the AG field in a typical race would be expected to have 300+ watt FTPs, since he himself had a 300+watt FTP and was only about top 10-15% on the bike in his races.

That's totally ludicrous - even in the races I've done where Strava has the "all-time KOMs" for the 1hr Oly segments, very few people do the entire thing in 300w.

Would you say then that the power held for an Olympic triathlon should be roughly similar to FTP (max 5% less)?

Most likely, yes. Although as peope have previously - FTP is a moving target, and the confounding factors on race day make it even more of a moving target, so I've learned to eyeball the number but not put blind faith in them.

In my best Oly PR bike split, I had trained up to a quite repeatable 230ish FTP based on 2 FTP 20' tests as well a lot more segments including the long Alpe du Zwift climb on Zwift. Went into the race with every intention of riding 220-230, as my prior best race power had been 215 (although I was in better shape since). Had a good big taper, and on race day, had a craptastic swim - I think it was the chop that threw me out of sync, and couldn't swim hard at all and underperformed it. However, once I hit the bike, thanks to the taper and the weak swim, I felt like a demon - dropped 247 watts for the entire bike split, which was higher than anything I'd ever done in training.

And before you say 'well, you overbiked' - nope. Proceeded to crush my run split, ran an Oly PR split that was a full 25 sec/mile faster than my prior year result, and which equaled my 800m interval paces in training - never thought I could hold that for a 10k, but I did. Ran so fast I almost equaled my racing buddy who is a primary runner and who trains nearly 1min/mile faster than me on average for all runs on Strava.

So count me in amongst those who don't look to the FTP as a holy grail. In fact, I've come to use my regular bike training loop numbers and segments a lot more than my "FTP".

No, I would say more like 85-95%. But again, picking a percentage like this is going to be highly impacted by small differences in definition/estimation of FTP.

In general though, if FTP is about the power you can hold for an hour and an Olympic bike leg is about an hour, I wouldn't expect to be able to do it at FTP after swimming and then still run well.

My n=1 is 20min test with no blowout lined up pretty much exactly with lactate testing in the lab. I'm skeptical 20min FTP test is that far out for the majority of people. Of course there will be outliers, but examples like the track sprinter in the linked article are obvious outliers at the extreme end of a normal distribution curve. If you are that concerned with 20min test accuracy, just do the full 1 hour. For most of us 20min test is a much easier option that gets us in the right ball park, and we can adjust from there.

Your 20 minute power lined up with the second lactate turnpoint power from a lab test? Or some percentage of your 20 minute power lined up?

Send one of these guys a note: Dr. Andy Coggan, Dr. Stephen McGregor, or Hunter Allen.

Dr. Coggan used to be highly active here and is very engaging. He could personally tell you the origin story of FTP and directly address your questions or concerns as a true expert on the concept. Anything else is second- or third-hand.

I've been quite happily using TR's ftp predictor recently as opposed to any test.

But I tend to think the entire testing debate is completely overblown. No test will be perfect unless one's in a lab and defining FTP as MLSS. And even then it'll be a different number the next day. There are probably people out there for whom x or y test is perfect or close to it. If a test gave a really good predictor for 60 percent of people, I'd call that a really good test, but it's still a lot of people for whom it doesn't work.

Assuming people are using FTP to set training zones, it's pretty easy to tell if it's wrong. If 2x20 (or whatever near threshold workout) is hard but doable on your average day, it's about right. If too hard, adjust down, if easy, adjust up. This is the part people screw up.

My maybe more controversial opinion is that it might be good for some folks to overestimate FTP -- if they just completely lack a base, and they would drop off a lot more than most between say 20 minutes and an hour, what they can do for an hour is maybe not such a relevant number for what might stimulate them best. But I'm not a coach and don't do that for myself, so no one need stress that I'm telling someone to do something really dumb.

I've certainly done my share of 20 minute test, but I'm more a fan of extrapolating from various rides what type of power myself or an athlete can achieve. Looking at many training rides provides more information than one 20 minute test does. I also like using CP, and plan to begin using a CP model with my athletes in 2023 to determine their power curve over various durations.

---

Blog: http://262toboylstonstreet.blogspot.com/
https://twitter.com/NateThomasTri
Coaching: https://bybtricoaching.com/ - accepting athletes for 2023