The Magical 4 Minutes
Just shy of 3000 words, estimated reading time: about 3½ Tabata.
If you’re like me your main information pipeline for fitness-related matters is Google web search. And there’s nothing wrong with that.
Even if Google Search profiles your preferences and you end up with more articles from T-Nation and Breaking Muscle than from
Strengtheory Stronger by Science. Provided that you check Google Scholar out from time to time, that is. Because as far as anyone can tell, Google Scholar does not care a rat’s ass about your browsing history with Google web search.
Google Scholar is my best friend and that brings me to the topic of the day: the Tabata protocol. This is something you definitely want to Google Scholar the shit out of. Here’s a common description of the protocol:
Tabata is the name of a Japanese researcher who discovered an interesting way to increase both anaerobic and aerobic pathways at the same time. […] It’s simple: take one exercise and perform it in the following manner:
- For twenty seconds, do as many repetitions as possible.
- Rest for ten seconds.
- Repeat seven more times!
That’s it! You’re done in four minutes! Oh, and that thing you’re trying to brush off your face? That would be the floor.
I pulled that quote from Dan John’s T-Nation post The Tabata Method (2004), and I’ll come back to it in conclusion. Other tidbits of information that you’d get from reading further are that those four minutes are as good as one hour of regular cardio and that most barbell and kettlebell exercises will do. Other sources will tell you that bodyweight exercises can do the trick, too.
All of the above is just fairytales and hokum. And it’s probably bullshit, too.
And there’s worse: I’m not the first to try to debunk the myth (see for instance here or here) but even the best-meaning debunkers seem to get it wrong. Why is that so? I have a couple of hypotheses but they are philosophical (read: probably untestable) so I’d rather keep them for another day because today is about facts.
When it comes to the original Tabata study, everybody seems to have alternative facts. That’s really sad because, unlike Trump’s inauguration pictures, finding the original study takes some digging. The good news is that once you read it, it’s easy to get the facts straight. Same as Trump’s inauguration, when you think of it.
And it’s worth digging because although our understanding of high-intensity interval training has progressed since 1996, all the addenda to the Tabata study’s conclusions have been incremental and the results still stand to this day.
So let’s get to business, and look at the study. If you have never checked it, prepare to have your mind blown. Because the Tabata protocol is not what you think.
The Myth of the magical 4 minutes
The table below summarizes what is usually said about the Tabata protocol and about what happened during the Tabata study beyond the 4-minutes-that-will-make-you-rub-the-floor-with-your-face-are-as-good-as-one-hour-of-regular-cardio thing.
|The claims||The Study|
|Test subjects were…||…both untrained subjects and highly trained subjects in different experiments.||Nope.|
|The participants were …||…split into 2 subgroups to compare the effect of moderate intensity and high-intensity exercise on aerobic and anaerobic fitness.||Slightly more complicated than that.|
|Exercise sessions comprised…||…either 60’ of continuous moderate intensity biking or four minutes of all-out biking (20 sec on, 10 sec off, repeated 8 times), 5x/Week||Significantly more complicated than that.|
|The conclusion was…||…that the aerobic group improved aerobic fitness, but not anaerobic fitness; the anaerobic group improved both.||That one is right. Yay!|
A full debunking of the Myth of the Magical 4 minutes would take a series of posts. Instead, I’ll complete the rightmost column for rows #1-#3 and put Row #4 in perspective. The first step towards this is to apply the #1 Analytic Fitness™ skill: Scholar-Googling the shit out of something.
Scholar-Googling the shit out of Tabata
Topical search is fine with Google Web Search, but it sucks with Google Scholar.
If you hit Google Scholar with “Tabata protocol” and if my assumption about the level of personalization of Google Scholar is correct, your search will look like mine, and the study won’t appear in the first 5 results or even in the first page. That’s bad.
But if you enter the list of co-authors of the original study in Google Scholar, you’ll get 2 hits, the first being the original study, and you’ll get a link to a free version. If you add the first names of the co-authors, you’ll get only one hit. Actually, I got down to one hit by just adding ‘Izumi’ before ‘Tabata’.
This approach assumes that you’ve already done some research on the topic, but it doesn’t have to be high-quality research either. For instance, I started with a post or two on Breakingmuscle.com and they puzzled me enough to go and check the reference. (In retrospect, as bad as these two posts were, they’re pretty much the best I’ve read.)
This approach can be turned into a general method for debunking science-based fitness bullshit:
- Don’t bother with Google web search and go directly to T-Nation and/or Breaking Muscle, then use their internal topical search.
- Check the reference you get (if there’s one), copy the list of co-authors, and Google Scholar it.
- If the study is not a landmark study, check how many citations the study has and click the ‘cited by’ link; look for meta-analysis, systematic reviews, and narrative reviews (also called ‘reviews’) referencing it, in that order.
From step #1, you’ll get a list of things to read for later for instance, if you want to get your certification in Analytic Fitness™ the hard way (cf. below for the easy way). From step #2, you’ll get the study and some additional benefits, like knowing what the co-authors have been up to since they published it. And from step #3, you’ll get a sense of how much you can trust the results (although the final test for that would be to check how the study is referenced in a textbook).
For the record, I just gave you a free lesson in Analytic Fitness™ and you can show your appreciation for it by supporting my Patreon. Some levels will reward you with an Analytic Fitness™ level 1 Certification without a sweat.
Or don’t. After all, it wouldn’t be a free lesson anymore if you did. Then again, it would still be a free lesson if you would show your appreciation for something else. Like, the fact that I put in the time to give free lessons. If you do appreciate that, you can show your appreciation by supporting my Patreon.
Checking the study out
The Tabata study is one of the most straightforward exercise science studies ever published. It has straightforward goals, protocols, and conclusions. I really encourage you to read it to get some further details. If there is enough demand from Patreon supporters, I may go back to the topic and squeeze more out of the study. But for now, let’s complete that 3rd column.
Row #1: Subjects
The study was done on 14 varsity athletes from different sports (soccer, baseball, rowing, etc.). The subjects were “moderately-trained young men” by the authors’ own lights. Neither untrained people, nor elite athletes. They were split into groups (7 + 7) so that one group would start with a slightly higher baseline aerobic fitness than the other (see Row #2: Groups).
In the discussion section, Tabata and his co-authors reference a study on untrained males that used a similar measure of anaerobic fitness but did not test aerobics fitness. Tabata et al. do not extrapolate from their conclusions to untrained subjects and comment only on the relative magnitude of improvements in anaerobic fitness in the two studies. This paragraph is probably the origin of the sub-myth that the Tabata study applies to untrained subjects.
As for elite athletes, Izumi Tabata has credited the idea of the protocol to Irisawa Koichi, the coach of the Japanese Olympic skating team. But Koichi was not involved in the study in any capacity. That’s probably the origin of the sub-myth that the Tabata study applies to elite athletes.
Row #2: Groups
Row #2 is where things really start to go wrong with popular science summaries. Mostly because explaining how aerobic and anaerobic fitness are measured is quite complicated. Fortunately, we do not need a detailed explanation.
The first thing to know is that, at the time of the study, there was a well-established measure for aerobic fitness (VO2max) but not for anaerobic fitness. To get around this, Tabata and his co-authors used a method to estimate the maximum oxygen debt that a subject can contract while exercising before total exhaustion (inability to move any further).
Exercising at the maximum sustainable oxygen debt exhausts a subject in 2-3 minutes of continuous effort. The ‘rest-pause’ protocol of 20’on-10’sec off was thus designed to allow for exercising 4′ at that pace, and (hopefully) train the anaerobic capacity.
Tabata himself had contributed to developing the method as a postdoc in Norway at the end of the 1980s. Establishing it as a standard for anaerobic fitness was one of the major contributions of the Tabata study. Getting into the further details of definitions and measurements would be extremely interesting, but it’s not necessary to debunk the myth. I’ll refer you to the original study if you are interested. Here are the juicy bits:
- Both initial aerobic fitness (VO2max) and anaerobic fitness were measured for all participants.
- The 7 participants with the highest VO2max constituted the Endurance Training (ET) group: they trained at 70% of their initial VO2max for 6 weeks, 5x/week, for 1 hour, on a stationary bike. Intensity was not adjusted during the 6 weeks.
- The 7 participants with the lowest VO2max constituted the Intermittent Training (IT) group: they trained 5x/week too. For the 20/10 intervals, the intensity was 100% of their initial anaerobic capacity, later adjusted for possible improvements (if they completed 8 rounds at aset pace, resistance was cranked up on the bike for the next session).
- At the end of the study, both aerobic fitness (VO2max) and anaerobic fitness were measured again for all participants.
Tabata et al. hypothesized that aerobic and anaerobic training is specific and that both groups would improve only in the quality that their training prioritized. Since the ET group did not do any anaerobic training, they expected no improvement in anaerobic capacity. And as expected, they found none.
They also did not expect significant improvement in aerobic capacity in the IT group but they found one. It was a major surprise but due to the design of the study, the specificity hypothesis was never contradicted.
Row #3: Exercise Sessions
This row is where the myth comes from.
Everybody seems to forget that the training protocol for the IT incorporated quite some aerobic training. About 3.5 times as much as anaerobic training, to be exact. Here’s what the IT group did:
- 4x/week, before their interval training, the IT group would warm up with 10 minutes of biking at 70% of their VO2max plus whatever biking was necessary to get there.
- 1x/week, the IT group performed 30 minutes of biking at 70% of their VO2max followed by an interval training session cut in half (4 intervals instead of 7-8) at an estimated 170% of their initial VO2max. This was never adjusted and therefore got easier throughout the 6 weeks.
So, instead of 5×60’=300’=5 hours of aerobic training vs. 5×4’=20’ of anaerobic training, what we have is 5×60’=5 hours of aerobic training vs. (410’+1×30’)=70’ of aerobic training + (4×4’+1×2’)=18’ of anaerobic training each week.
This explains why specificity was never violated: the IT group performed 23-25% of the aerobic training of the ET group: (20’/300’)=.23, rounded up to .25 to account for the few minutes every session needed to climb up to 70% of the V02max.
So, what conclusions can be drawn from that? Well, here we go.
Row 4: Study Conclusions
This is the one everybody gets right. Or wrong, depending on how you look at it.
- Everybody gets right that the IT group improved both aerobic and anaerobic fitness while ET group improve only aerobic fitness.
- Most people get wrong how and why it is so. In fact, based on what they read, most people believe that anaerobic training violates specificity.
And in fact, it does not. If they were the case the Tabata study would have been a revolution. It was not. It was a major breakthrough, but not a revolution. What, then, was that breakthrough?
Well, it’s what researchers in the field would call a dose-response finding.
What Tabata and his co-authors established is this: if you take the ‘dose’ of cardio of the ET group and their ‘response’ to that dose (their improvement in VO2max) as reference, then ¼ of the reference dose of cardio produces more than ¼ of the reference response when supplemented by a modest dose of maximal anaerobic training.
And they could not really say more than that because they had not expected it. For instance, they could not draw further conclusions from the observed magnitude of the improvement of VO2max in the ET and IT group because they could not know what had caused the latter. But before I get to that, let’ spin things around a bit.
The importance of the frame of reference
Going from zero cardio to 1 hour and 15 minutes is a pretty big step. Still, with the right frame of reference, the real Tabata protocol does not look that bad.
Especially with the right choice of colors. For instance, the graphs below pit the expectations one could have based on popular expositions (on the left) against the reality (on the right), with cardio represented in red (because it sucks) and interval training in green (because it’s cool).
What the picture doesn’t tell you is what’s red and green stand for. 70% of your VO2 max (red) would get you sweaty, 100% of the VO2max would make you nauseous, and maximal anaerobic capacity, which averages 170% of your VO2 max (green) would first make you feel queasy, then puke, and then cause your muscles to shut down (after 2-3 minutes). That’s what green stands for: an intensity so high that your limbs would go limp if you tried to maintain it more than 2 or 3 minutes.
Left on our own, we would never reach that kind of intensity. Humans used to have predators, and trying to outrun a predator at a pace that would make your body shut down after 2-3 minutes is not very smart. It makes sense that we’d have some advance warning (queasiness).
That’s why Tabata and his colleagues chose mechanically-braked stationary bikes and put a lab assistant to ‘encourage’ participants to keep the pace and stop them if they did not. When you bike against a resistance at maximum intensity, the limbs responsible for motion never rest: your leg extensors on one side are assisted by your leg flexors on the other and when they finally relax, the leg flexor begin to assist the extensors on the other side that just kicked in.
(Also, for later: with bodyweight, barbell, and kettlebell exercises, the limbs responsible for the motion alternate between loading and unloading. Make a mental note of that, I’ll come back to it in conclusion. With a vengeance.)
Now, with and intensity of 70% of your VO2max, the leg extensors on one side do not need the assistance of the leg flexors on the other. Once you get into the rhythm you can actually maintain it while doing something else, like watching Netflix or reading a book. So we might as well represent the situation by switching green and red around and putting a catchy title. Suddenly, the ET protocol becomes more appealing.
For the number-crunchers
The previous section showed the importance of spin. With this one, we go back to science, and the importance of getting the numbers right.
If you read the study and crunch the numbers, you’ll realize that the IT group improved their VO2max 1.4x more than the ET group, so it may seem that the IT group got a whopping 140% response for a measly <25% dose. Even without violation of specificity, that could have started a revolution in endurance training on its own.
But it did not, for two simple reasons. First, the ET group started with a higher VO2max, and therefore closer to their maximum potential. Thus, any dose that would produce any response at all would have been expected to produce a higher response in the IT group.
Second, the level of effort of the ET group was never adjusted. And yet, their VO2max steadily improved. There is no way to tell how much they VO2max would have improved had their level of effort been adjusted in the same way as the level of anaerobic effort of the IT group was.
Again, the surprise came from the IT group improving their aerobic fitness, not the magnitude of the improvement. Since Tabata and his co-authors had not anticipated the improvement, they hadn’t taken any precaution to compare the magnitudes of improvements in VO2max across groups in a meaningful (=statistical) way. And, as any good scientist, they refrained from saying anything else about it than stating the figures.
Conclusion: Tabata™, bullshit or not?
Let’s go back to my quote of Dan John’s T-Nation post, The Tabata Method (2004). As far as I’m concerned, Dan John is God. He’d probably object to the description because he’s a devout Christian, but I’m an atheist so I can choose the Gods I disbelieve in.
But when Dan John recommends that you do ‘Tabata intervals’ with kettlebell goblet squats, kettlebell swings, and farmer’s walks, I’m that close to calling bullshit. There’s no way anyone could reach maximal anaerobic capacity in 20 seconds with these exercises. There is just too much loading/unloading going on.
So why does Dan John recommend them, then? I don’t know, but he has vested interests in RKC (®, ™, or whatever) and makes good money out of it, and those are exercises that he teaches in RKC certification seminars. That could be an honest mistake or a personal bias. That could also be bullshitting.
Here’s another nugget of gold: there’s no way anyone could reach maximal anaerobic capacity in 20 seconds with bodyweight exercises either. So, armed with that fact, take a look at that.