Interval Walking Training: what the Japanese 3-and-3 protocol really is

The prescription

The IWT protocol, block by block

Nemoto 2007 words it as “five or more sets of 3-minute low-intensity walking at 40% of peak aerobic capacity for walking followed by a 3-minute high-intensity walking above 70%”, done four or more days per week. Ten blocks, thirty minutes, no equipment.

1. 0:00 Fast block 1 above 70% of walking VO₂ peak
2. 3:00 Slow block 1 about 40%
3. 6:00 Fast block 2 above 70% of walking VO₂ peak
4. 9:00 Slow block 2 about 40%
5. 12:00 Fast block 3 above 70% of walking VO₂ peak
6. 15:00 Slow block 3 about 40%
7. 18:00 Fast block 4 above 70% of walking VO₂ peak
8. 21:00 Slow block 4 about 40%
9. 24:00 Fast block 5 above 70% of walking VO₂ peak
10. 27:00 Slow block 5 about 40%

Block length

3 minutes, both kinds

Repeats

5 or more, so 30 minutes and up

Frequency

4 or more days a week

Fast minutes per week

60, at 5 repeats × 4 days

That last row is worth a moment. Five fast blocks four times a week comes to 60 minutes of fast walking a week, and the 2019 dose-response paper found benefits climbing with fast walking time up to about 50 minutes a week before flattening out. The original prescription lands just past the knee of that curve. It was not arbitrary, and doing double does not buy you double.

Origins

Where Japanese interval walking came from

The method belongs to Prof. Hiroshi Nose and Dr. Shizue Masuki and their group at Shinshu University Graduate School of Medicine, in Matsumoto, Nagano. They have been publishing on it for close to two decades, mostly in adults in their 50s, 60s and 70s, and mostly at a scale that is unusual for exercise research.

The landmark trial is Nemoto et al. 2007 in Mayo Clinic Proceedings. Two hundred and forty-six people, mean age 63, randomised for five months into no walking, moderate continuous walking, or interval walking. The interval group came out ahead on leg strength, on aerobic capacity and on resting systolic blood pressure. The 2011 cohort study in the British Journal of Sports Medicine took it out of the lab and into the field with 826 participants, and found the biggest improvements in the people who arrived least fit.

One small thing that most write-ups get muddled, since it costs nothing to get right: the first author of the 2007 paper is Ken-ichi Nemoto. Mayuko Morikawa is a different researcher in the same group, and she first-authored the 2011 cohort paper. They are two people, not one.

The result has also been reproduced outside Japan. Karstoft and colleagues in Denmark ran a four-month randomised trial in adults with type 2 diabetes where the interval and continuous arms burned the same energy. Only the interval walkers improved their VO₂max, body composition and glucose control. Small trial, particular population, but it is an independent group getting the same shape of answer.

The evidence

What the trials actually measured

Five studies do most of the work. Every number below is from the paper it sits next to, and every card links to the paper so you can check us.

Two caveats that belong right here. Most of this work comes from one research group, which is a real limitation no matter how large the cohorts are. And the gains above are changes measured against each group's own starting point, in people who were mostly sedentary when they began. If you already train hard, this literature is not about you.

Corrections

Five things the internet gets wrong about interval walking training

Japanese interval walking had a viral year, and a lot of what got written during it is downstream of other articles rather than of the papers. These are the errors we ran into most often.

1. “95% of people stick with it.”

   No paper in this literature reports anything close. The only long-term adherence study, Masuki 2015, found 70% over 22 months. Nemoto 2007, which is where the “mean age 63” detail comes from, reports no adherence rate at all — what it does report is that the prescribed targets were met by 11 of 19 men and 31 of 68 women. A 2024 review co-authored by Nose and Masuki says plainly that long-term adherence “remains a challenge”, especially for people who are overweight or managing a chronic condition. We had this wrong on this site too, until we went and checked.

2. “A ten-year study proved the benefits last.”

   There is a ten-year item, but it is a conference abstract in a FASEB Journal meeting supplement, not a peer-reviewed paper. A sweep of every interval walking paper in PubMed authored by Nose or Masuki turns up no full-length ten-year follow-up. Longest published follow-up we can point at is the 22 months in Masuki 2015.

3. “The protocol starts with the fast block.”

   Nemoto 2007 words the prescription the other way round: five or more sets of three minutes slow followed by three minutes fast. It is a repeating cycle, so in practice the two orderings are the same session, and starting fast after a warm-up is perfectly reasonable. But the paper does not prescribe fast-first, and articles that say it does have not read it.

4. “It works because you cover more ground.”

   The 2019 dose-response analysis is specific about this. Fast walking time predicted the gains. Slow walking time predicted nothing. Total walking time predicted nothing. The slow blocks exist so you can do the fast blocks properly, not because the distance adds up.

5. “Interval walking lowers your blood pressure.”

   On average, across a group, it does — that is the Nemoto 2007 result. But Masuki 2010 in Hypertension found the blood-pressure response varies with vasopressin V1a receptor genotype. The effect is real and it is unevenly distributed. Your mileage genuinely may vary, and anyone stating a single number for what it will do to your reading is overselling.

Intensity

How hard is “brisk”, exactly?

This is the part that decides whether the protocol works for you, and it is the part almost every summary hand-waves.

The trials did not define the fast blocks as a pace in kilometres per hour or as a step count. They defined them as a share of your own peak aerobic capacity, measured in a graded walking test: above 70% for the fast blocks, around 40% for the slow ones. That is a per-person target, which is exactly why it is hard to copy from an article. Someone else's brisk is not yours.

Nothing you can wear measures oxygen uptake, so any wearable implementation is using heart rate as a stand-in. That substitution is a practical approximation and worth naming as one. If you are going to approximate, approximate well: heart-rate reserve, which scales the range between your resting and maximum heart rate, is individual to you. An age-based formula is a population average wearing your name, and it can miss a specific person badly.

The zero-equipment version is the talk test, and it is better than it sounds. Fast means you can still get words out but full sentences are a struggle. Slow means you could hold a conversation comfortably. If you can sing, you are not pushing. If you cannot speak at all, you have overshot the protocol, which is brisk walking, not a sprint.

Expect your heart rate to lag the switch by twenty or thirty seconds in both directions. That is normal physiology, not a failure to try hard enough. Judge a block by how it felt in the middle, not by what the number said in the first few steps.

No app required

How to do the IWT protocol with nothing but a watch face

You can start this afternoon, for free, with whatever is already on your wrist. Here is the honest version.

1. Pick a route without traffic lights. Interruptions at minute two of a fast block are the single most annoying way to ruin a session. A park loop, a track, a long quiet street.
2. Walk easy for three to five minutes first. The protocol as published has no separate warm-up, because the slow blocks do that job. Starting on a fast block cold is your call, and a short walk-in makes the first one honest.
3. Watch the clock and switch every three minutes. Ten blocks. Most people find the third and fourth fast blocks are where it gets real.
4. Push properly. Undercooking the fast blocks is the usual failure and, per the 2019 paper, fast time is the thing that predicts the result. A tepid fast block is a slow block that took more effort to talk yourself into.
5. Four days a week, and count weeks not sessions. The trials ran five months. The one that measured adherence ran 22. Nothing here is a fortnight-long project.

The friction is real, and it is specific: glancing at a phone or a watch face every three minutes for half an hour, in the rain, while trying to keep your effort up. That is the problem worth solving, and it is smaller than most fitness apps pretend. If a kitchen timer and a repeating alarm work for you, use those. They are free and the protocol does not care.

Fit

Who it suits, and who should ask a doctor first

The population studied here is middle-aged and older adults, mostly sedentary at baseline, with a lot of women in the cohorts. If that describes you, this is unusually well-matched evidence. It is low impact, it needs no gym, no equipment and no skill, and the 2011 field study suggests the less fit you start, the more there is to gain.

It is a poor fit if you want something you can do in fifteen minutes, or if you already train at intensity, where this is likely to be easy aerobic work rather than a stimulus.

None of this is medical advice. Brisk walking raises your heart rate on purpose. If you have a heart condition, uncontrolled blood pressure, joint problems, or you have been away from exercise for a long stretch, talk to your doctor before you start. The 2024 review notes that people with chronic disease are also the ones who find long-term adherence hardest, which is worth planning around rather than being surprised by.