Empirical Cycling Podcast “Why Not Rønnestad 30/15 Intervals” aka Brasted

In the most recent podcast “Watts Doc #23” (21st July) Kolie is highly critical of the Rønnestad study (from about the 1hr mark). He identifies what he sees as a number of flaws in the methodology leading to some “disingenuous” conclusions being presented by the study. In a nutshell he argues that whilst the 30/15 interval format studied (embodied in the TR Brasted workout) is capable of delivering useful improvements in anaerobic power it will not improve long term VO2max.

The study in question can be found here: https://pubmed.ncbi.nlm.nih.gov/31977120/

Now as someone who has adopted Brasted as my go to VO2max workout I felt I was in good company when coach Chad mentioned in a recent podcast that he was also favouring this format. Can anyone with a better understanding of the physiology than me muster a defence for Brasted or am going to have find an alternative (and likely less pleasant) VO2max workout?

EDIT: You really will need to listen to the relevant part of the podcast to understand kolie’s position - my very brief summary was just to give you a flavour of its content for you to decide whether or not you were interested enough to do so.

Also just edited the post to add the qualifier “long term” in the comment about VO2 max improvements in order to reflect (hopefully correctly) a comment made by Kolie in a post he has added below.

1 Like

Neither Coach Chad or Kolie are PhDs doing research, so we should understand their comments and attitudes towards studies as more of a “expert coach” then an “expert researcher”. They might not understand everything in there, but at the same time, they can provide context and experience that a researcher does not have.

Science is sloppy, people make mistakes and one study does not a field of research make. I haven’t listened to this particular podcast but I do pause when I see laymen accusing scientists of being disingenous or somehow purposefully ginning the game. I’m not saying it doesn’t happen and that peer review is bullet proof, but I think most working researchers are too mindful of their reputation see a benefit in producing one-off gin-up results. A lifetime of work can be thrown away with one attempt to force an issue.


Have a listen to this. https://scientifictriathlon.com/tts243/. In one place he says that certain types of HIIT intervals raise MAP but not VO2max implying they are more effective at increasing economy.


I thought the short interval protocol was originated in earlier papers such as this: https://pubmed.ncbi.nlm.nih.gov/24382021/ or even the 2013 predecessor to this.

Now, I am just a layperson and very interested to listen to and digest the empirical cycling criticism at that. But at least the claim made in the OP that 30/15s fail to raise vo2max but affects anaerobic power is not supported by the 2015 study - every single subject doing the short intervals raised their o2 consumption per fig 3 a. Hence the interest towards the protocol is quite understandable.

Perhaps there are methodolohgical issues that render these data less representative or reliable. Nevertheless, the data is there. Perhaps the 2020 study is different. Or perhaps the claim I commented on was just OP’s way of paraphrasing the podcast.

Either way, this should be an interesting listen!

Thanks Buzzcock. For sure I am paraphrasing, and you are quite right to be cautious of putting too much weight on that - especially given my limited understanding of the science! I would definitely recommend you give the original a listen.

I have just gotten out of bed and won’t have time to re-listen to the podcast for a while. There is obviously a lot more in Kolie’s analysis than in my summary - I really couldn’t do it justice if a tried to expand and it has to be heard first hand. But a couple of things I do clearly recall are that Kolie definitely attributed the measured improvements in 20min power to improvements in the anaerobic power component. He also mentioned that reductions in measured lactate concentrations were more likely due to improvements in lactate shuttling (and also maybe increases in blood volume) than reductions in lactate production.

I’ll stop there because the more detail I try and report the more chance I will get something horribly wrong.


Quite the contrary, I think you opened an interesting and informed discussion that probably induces learning amongst participants.

Will give it a listen when time permits!


Please someone correct me if I misunderstood, but I think he is suggesting the athletes in the long interval group were probably at best riding at FTP and not at VO2Max. The intensity was matched on RPE alone.

1 Like

Buzzcock: I forgot to mention that whilst Kolie’s comments were directed to the most recent study he did refer to the earlier study that you linked and said that it fell into the same boat (my words). I believe the distinction was that the most recent study was done on more highly trained cyclists than the first.

So even if my HR reaches 90-95% of max for a significant period of time that has no impact on VO2 max?

This may or may not be the case here, but feels like with the abundance of internet based platforms, blogs, podcasts etc everyone is just trying to make a name for themselves now, usually by something semi controversial and typically without hard science to back it up. Becoming too hard to track what should be believed unless you live this stuff.


I’ll have to listen to the podcast again, but I think the gist was more that:

  1. The comparison in the study was of 30/15s @ VO2 max power (I know, I know, not a thing…) to 5x5 at c.FTP. This comparison flatters the effectiveness of 30/15s at increasing VO2 max.

  2. I don’t recall the claim being made that 30/15s don’t improve VO2 max, more that they aren’t the most effective, and that what gains are made plateau sooner than alternative methods.

My own personal experience with 30/15s was that they didn’t do it for me. I’d wind up being able to perform better on a ramp test, but then fail when I tried a TT effort. Unfortunate as I ‘enjoy’ the workout protocol itself.

I actually tried to implement the tips he gave in the podcast in a session last night - I’ll probably post my thoughts on them in the other thread.


I was referring to the deep dives in Watts Doc podcasts 18-22. Was often left in the dust as I have no specific background in biology.

Episode #23 was an easy listen, with actionable takeaways.

1 Like

@Peter you are correct, I’m not employed as a scientist doing research. I studied biochemistry and physiology as an undergrad and have continued my education on my own, but can tell you for a certainty the podcasts are assigned for high school, college, and masters/phd education. I agree completely that peer reviewed science isn’t bulletproof, as the scrutiny of peers after publication is the real peer review. My ability to validly critique the literature comes from a somewhat rare but definitely not unique position of being a scientist by training, but coach as well. Researchers are indeed mindful of their reputations, but some are blind to what effective training is. Ivory tower syndrome.

My critique of the Ronnestad 2020 paper comes from the interval comparison where 4x5min at about tempo/sweetspot/ftp (depending on the athlete) is not enough to qualify as vo2max training, and definitely insufficient to even qualify as effective FTP intervals. @iMatt66 does a good job summarizing the salient points, and I’ll add one more he missed, which is that higher power intervals will use and train larger motor units as well. So overall, what Ronnestad did was put Thor Bjornsson in a wrestling match against a crash test dummy, whether he knew it or not.

Overall, yes, the 30/15s group did increase vo2max. But let’s also put a short term improvement in vo2max in context of the preceding 5 podcasts, where we show there are different mechanisms of improvement in the short, medium, and long term. My interval suggestions are entirely focused on the long term. One of the FTP training studies we used showed an increase in vo2max from FTP intervals. That would be a short/medium term adaptation that obviously does not last forever. Exploring what’s “under the hood” of a measurement like vo2 is the point of the last five episodes. It’s something that most exercise physiologists (except a few of the modern day greats like Montero and Lundby) don’t try to do in their studies. It’s like saying “this car has X more horsepower”, we don’t know if it got forced induction, maybe it revs higher, did it get a cat delete, more cam lift, electric motors, who knows.


I don’t want to distract from the conclusions, and thanks again for the great podcast, but I wonder about the 4×5 group. You think they worked at too low a power, probably under ftp. But their mean work intervals were at 368W, and their mean 20min power was 348W. Doesn’t that mean it was likely above ftp, irrespective of the result of the ramp test?

I actually think this is a side issue, when the real point if discussion should be the improvement of vo2max vs anaerobic power, but I don’t think the study sheds much light on that, partly due to the non-improvement in the 4×5 group.


Good catch! I love this so much. You’re right in that the discussion should be vo2max vs anaerobic power, as we mentioned about the original Tabata study. But you bring up something that got left on the cutting room floor.

Here’s the long version about work rates and 20min test and whatnot that I cut out of the podcast, since it weaved in and out of a tangential discussion about getting into averages vs individualizing datapoints, RPE, instructions and compliance, and how we can see conflicting data points resolve when we look at individual points. We might include it as an addendum in the next episode since it probably does need to be addressed.

79+/-7% wmax interval avg work rate average likely shows a heavy left skew and sure, some working over FTP but most not. Here’s how I figured that. Looking at individual responses, we can probably say three were working over FTP who did show a positive response, the rest were not, or I’m absolutely certain they would have shown a bump in vo2max. But this is a problem covered up by the averaging so we’re left with educated guesses.

We can try to use power at 4mmol to approximate FTPs, so we have a conflicting data point in the study where it’s about 320-330w and yet 368w is only 79% of Wmax, what we might estimate as vo2max power. So by this data and knowing most elites are at FTP around 80% vo2max but almost definitely more like 85% or more. 330w around 70% of Wmax (or estimating threshold around 70% vo2max) is REALLY low for an elite endurance athlete. So 4x5 is either working above threshold by a couple watts, or below threshold by a couple % vo2max. Conflicting data point brought about by the methodology. If all of them were working over threshold, I’m sure we’d have seen the same (as 30/15s) or better response in vo2max in the 4x5 group. But this gets resolved.

I’m sure 4mmol power (general population average for MLSS concentration) underestimates FTP in a ramp test by a lot, which sounds a little like heresy at first, but we just can’t use this to estimate threshold and here’s why.

As you begin to recruit new motor units as power output goes up, they start with a bump in lactate as glycolytic flux “covers the over” for a few minutes and then settles down. I’ve seen countless ramp tests that show 4mmol, or just plain inflection point as determined by Dmax or whatever other method you want to use, at 30-50w below FTP. Someone who does this right is San Millan, who looks at blood lactate in the right way, having people do a 10 minute steady power output and measures blood lactate only at the end. Just opening up a random ramp test with lactate data, looks like this guy’s FTP by blood lactate ramp test is about 300-310w, but in reality is 360-370w. So without letting the fresh motor units get used to their new duties, blood lactate is way higher than it’d be without a period of settling in.

20min test: this followed the ramp test after 20 minutes rest. Any coach who’s ever had folks do the 5min blowout before 20min test knows that this doesn’t increase 20min test accuracy at all, it reduces it to the point of uselessness. Some folks can barely do FTP or SST watts at that point, so I had dismissed it in this study as a valid test to estimate FTP, or even maximal 20min power. The point about increasing anaerobic and buffering capacity, better trained larger motor units still holds up here. Anyway, that’s why for the 20min test the groups averaged 5-7mmol to start.

The protocol says they aimed for highest average power output “for each session”, but this is part of the RPE problem, and problem with instructions. The Borg scale is so weird anyway, it’s out of 20 and there’s very little granularity with how we describe RPE on the bike, which is why I’ve never used it to describe anything to athletes. It’s more like for people who don’t really exercise… Borg’s 19-20 is something “you can’t maintain for long”. How long is long? Does it have to be a sprint? So okay, now we have 14 is basically endurance pace (not out of breath), and 19-20 is a sprint or damn near. Not much room to work with. VO2max intervals usually rate an 8-10 for my athletes, FTP 7-9 depending on duration… but those all feel very, very different from each other regardless, so there wasn’t great control of actual work intensities, and whether or not power values wandered at all doing 4x5s and figured some fatigue came into play, but either way they were definitely not working that hard regardless of the instructions.

This is why I went with the estimation method I did, and made a rough assumption, plus the individual responses, that most of the 4x5min intervals were generally at or below FTP, and the distribution of actual 5min interval power outputs skewed low.


Sorry for my uneducated question - so in that case what is the point of doing 20 min power test? To fill the power curve? If you use long protocol to test the ftp, 20 min power becomes only the pont in the power curve and can be useful value during races?

Just an FYI: This recent study reported that the warm-up has no effect.

Another FYI: Borg’s 6-20 scale is meant to correspond to typical resting (60) and maximal (200) HR. All of the action is therefore at the high end because many responses other than HR aren’t linear with intensity.

Also: A.V. Hill didn’t invent ramp testing - in fact, he didn’t even use ramp tests, instead relying on a discontinuous protocol.

1 Like

What warm-up has to 5 min effort before 20 min test?

The 20 minute power test can be used to estimate FTP, somewhere in the range of 89-90% to 96% of the 20 minute value. Hunter Allen says it’s easier than doing a long TT, but I disagree. Either way, when I have someone do a mid-range test, it’s usually 12-25 minutes, but just to keep the wko5 model up to date if I’m using it to track fitness changes. I would be cautious about extrapolating the meanmax power curve too much in races. What you can do after being fatigued changes, and (hopefully) improves as you become fit.

1 Like

Thank you for your response, and I agree that I have no Idea how anybody can think that 20 min test is easier than long FTP test.

1 Like

Having done a PhD myself and working as an active scientist, I appreciate this comment a lot.

Of course there are flaws in studies and in hindsight one might have done a detail different here and there.

Essentially, in hindsight, everything is easy to criticize. Good scientists live with this, judge their data appropriately, participate in discussions on uncertainty/ambivalence, and improve next time around.

Thus, there are always parts in the methods/setting that can be discussed. Some people even have completely different takes on interpreting the same results.

But the data from such studies give us the base to have these kind of discussions in an informed way, with the chance to create something new/better in the future.