Why basing vo2 off percentages is misguided

The following was shared with me by EC coach and donut enthusiast Rory Porteous, who is currently training in Mallorca.

the first 13mins, according to him, was done at 121%. He also tested FTP 4 days ago and his ftp is 264. Now, this may be an extreme example but it shows that sticking to percentages is really not the way to go when thinking about vo2 work. On the other side of the coin, I struggle with hitting 120% on 3min efforts and my 5min power is around 112% in recent workouts. It took me a long time to come around on this, and I hope by sharing this maybe we can get a few more people off the vo2 % of FTP train


The Empirical Cycling Podcast guys get fired up about this same thing.

Personally, I try to do all my VO2 max work outdoors and if I do have to do it on a trainer I do it in resistance mode not ERG.

I think VO2 intervals need to essentially be the highest power I can reliably maintain over the duration (factoring in repeats and rest) rather than being based off a fixed % of FTP.

For short intervals less than 2 minutes my power needs to be way over my TR % of FTP target to provide a proper stimulus. Once you get into the higher progression levels and start getting some 5+ minute VO2 intervals I’ve found my repeatable productive power is pretty close, but still slightly over my prescribe TR power based on % of FTP.


This has been known for a while now, wko5 added iLevels to help combat this issue and good coaches don’t use set percentages above ftp.

Not sure why TR still uses a set % rather than adjusting it based on the athlete though.


You learn more from your failures. He said “full gas” and it was my first vo2max that build…

overcooked the first one at 133% ftp, then 126%, then 120%, and on the 4th mentally surrendered to my screaming legs.

One of the first hot days, did those around 10am it was already 82F / 28C. Good times.


Haha my first vo2 workout is almost always a bit off. First interval overly ambitious then just hanging on for as long as I can.

After that though they are dialed in, still feel like death but at least get my target time in.


Totally agree, but am I wrong to think PLs approximately address this?


I just think we get hung up on the method. Do whatever works.

1 Like

Probably because for a piece of software to do this on an individual basis would be complicated in terms of the workout output. You can easily tell someone max repeatable effort and coaches do this, but when creating a specific workout file, you need some value to set it to, whether right or not. Also, max repeatable effort can be tricky to figure out when new to this, so %FTP can be helpful as a place to start, but as one becomes more experienced, I think moving away from that is helpful.

FWIW, I’m fully on the VO2 intervals in resistance mode and do them at a max repeatable intensity regardless of what a prescribed TR workout may say for intensity. And sometimes that backfires a bit when I overcook the first interval in the first workout, but then I can adjust in future workouts. I prefer that over trying to hit a set power for each interval that may or may not be right and doesn’t allow for power to drop a bit in later intervals.


They both work one is just better than the other. I would personally rather do a more beneficial workout right away than take a few iterations waiting on a system to dial it in.

Sure it will hopefully get me in the right place eventually and even the failed workouts probably have some training effect but if one has limited time it makes the most sense to go with the better setup from the start.

1 Like

Yep, it’s a drawback they seem to be trying to address with progression levels.

Not sure how much they do under the hood with an athletes PDC but I have found, if it’s recent and maximal, you can get a very very good idea of where someone stands just using that.

1 Like

Before adaptive training I saw two reasons:

  • the popularity of Erg, and TR making Erg the default
  • the use of Coggan Classic zones

Adaptive Training changed things by creating a family of workouts with different % FTP. So in theory for vo2max workouts you can get a higher/lower % FTP by choosing or letting AT choose workouts with a higher/lower vo2max difficulty rating.


A lot of structured training is based on standardization - it has been done this way, so keep doing it this way. I personally think you’re better off determining your all-out, one-time power for a certain duration then determining what percent of that all-out effort you need to do if you want to do it X times. So if your all-out, single 5 minute effort is 400 watts (and you literally cannot do any more intervals), start with 90% or 92% or whatever you judge appropriate and see if you can complete however many intervals you plan to do. If you can complete them then use RPE to determine if you should add 5 or 10 watts or keep it as-is. If you cannot complete them cut 5 or 10 watts, see how it goes. You’re still getting a good workout and you’re learning to feel the efforts, what you can handle and what you cannot, which is really important. Shouldn’t only go with numbers on a screen.


thats what I do, go out and do a max effort. Then pick a % of max and adjust as necessary.

Nope I don’t believe PL are appropriate for vo2 and I hardly think they’re valuable for endurance work either.

1 Like

On the podcast, both @Jonathan and @Nate_Pearson have said it makes sense to do VO2 max and anaerobic/sprint efforts in resistance. You don’t have to be wedded to the %.

But honestly, for a workout in TR, how else would you design it if you didn’t use % as a general mark. To me I can see using erg if it’s something like 105%-110% and a longer 4-5 minute effort.

I generally do the first set on erg, see how that feels and then switch to resistance. But TR isn’t making you do it on ERG.


If it were my trainer control app, the app would show power ranges and not some exact power number. And then you could have workouts like this with a power range, and if ISO power (repeatable power) the app would give instructions to go hard but repeatable without too much drop off, and that finding a good target can take some trial and error, so don’t get discouraged and think you failed if you go too hard and power drops a lot. The app would turn off Erg and putting you in resistance. And of course that means having a way of helping the user set a good resistance % for that target power range. The app would also have a mode to estimate your anaerobic capacity and ftp, and then a week or two before doing vo2max work have you do an all-out effort to figure out the “curve under the curve” and suggest target power range and take into consideration your anaerobic capacity.

I literally do that myself, it’s pretty easy to design intervals and estimate burn down of anaerobic capacity.

The use case is “just go as hard as you can and forget repeatable (ISO power)” which is a little easier because you just go all-out and it ends up where it ends up.

Just thinking out loud, I can see what decisions TR made which appear to be driven by satisfying the Erg users, and a reluctance to embrace power ranges.

1 Like

Surely it just needs to know the power profile of the athlete. Then it could be parameterised value rather than just being a fixed percentage of another part of the power curve.

This is fine if people have done maximal or near-maximal efforts at appropriate durations, either recently or at all.


If doing VO2 max appropriately the right range will come out.

1 Like

Interesting data!

Of course, this is old news on this forum. It’s been said on this forum for years: VO2max is not a power number. Rory is saddled with the same problem many athletes are & that is poor fractional utilization. Sadly, Rory is a ‘Group L’ cyclist (me, too, btw!) in the parlance of the classic Determinants of Performance.

Also found on this forum is how to fix that problem! But if you want to ‘go to the source’ you can see Gollnick’s classic “Effect of training on enzyme activity and fiber”. Or just read the threads. ;-D Although even I will admit it can be thick reading.