Clay Tech

"clay-works make things real"

translated from clazytech.com

No one knows how many 2-meter-tall Japanese people exist

original Science

Not a single one among my acquaintances

Last month I was watching the volleyball Nations League. Japan’s national team is small compared to other countries, but even so, several players over 200cm show up.

That’s when it hit me. Among the people I could call an acquaintance, there isn’t a single Japanese person who is 2 meters tall.

At 190cm, there are a few. In the 180s, it’s not even unusual. But at 2 meters, it’s zero. They only exist on television.

So how many are there, actually? I decided to look it up.

When I tried to count, there was no statistic

Let me state the conclusion first: it doesn’t exist.

Japan has only two height statistics. The National Health and Nutrition Survey, and the School Health Statistics Survey.

The National Health and Nutrition Survey has a small sample. In the 2018 physical condition survey, the number of men whose height was measured, summed across all ages, was 2,900. To capture a 1-in-100,000 event, you need three more orders of magnitude than that. The published format is basically means and SDs, and even the detailed tabulations stop at the 99th percentile. In the 2016 survey, men aged 18-29 came in at 184.9cm. That’s where it cuts off.

The School Health Statistics Survey has a sample of tens of thousands, but it covers ages 5 through 17. No adults included. And the distribution tables are published in per-mille (‰), so the resolution floor is 1 in 10,000. A 1-in-100,000 event falls below the display unit.

The only census-scale data comes from the specific health checkups (tokutei kenshin). These actually measure about 30 million people a year, and height is recorded too, in order to compute BMI. But the published distribution tables cover BMI, not height. And the checkups are limited to ages 40-74, which drops the tallest cohort — people in their 20s and 30s — entirely.

Historically, the only height census Japan ever ran was the military conscription physical. It measured nearly every 20-year-old male. But the average back then was around 160cm, and the units were coarse. Unusable.

Forced to estimate, and the estimate itself broke

With no statistics available, the only option left is to calculate from a normal distribution.

Mean for adult men, 171.5cm; SD, 5.8cm. 200cm is 4.9σ. The upper-tail probability is 4.5×10⁻⁷, which, across 50 million adult men, comes to around 20 people.

But this number turns out to be completely unreliable.

Set the SD to 5.5 and you get 6 people. Set it to 6.0 and you get 70. Move a single parameter by 0.2 and the answer changes by an order of magnitude. Beyond 4σ, the estimate simply doesn’t hold up.

There’s a worse problem still. At this height, disease-driven cases start to outnumber cases of ordinary individual variation. The prevalence of Marfan syndrome is about 1 in 5,000. Growth-hormone-secreting pituitary disorders are designated as intractable diseases. A statistical problem — “average + 5σ” — turns partway through into a prevalence problem.

Not only is it impossible to count, the estimate itself breaks down.

Using sports as a sanity check

So I tried proof by contradiction instead.

Add up the V.League, B.League, and sumo, and there are several dozen Japanese athletes over 200cm. If the underlying population were only 20 people, every single one of them, nationwide, would have to have become a top athlete. That’s obviously off.

In other words, the normal-distribution estimate runs too low. The real figure is probably in the hundreds — somewhere around 500 to 1,000 seems reasonable. That works out to roughly 1 in 100,000 to 200,000.

Even if the number of people whose face and name you can match over a lifetime is a few thousand, the probability of encountering one is a few percent. No wonder none are among my acquaintances. Incidentally, at 190cm the rate is 1 in 1,400, giving an expected value of 1 to 2 people. My gut sense of “a few” turns out to be roughly right.

The reason it can’t be counted and the reason everyone gets found are the same

Here’s something I noticed.

The world of sports, which I brought in as a sanity check, turns out to be the more interesting part.

There are only 500 to 1,000 such people, and yet several dozen of them play in top leagues. As a utilization rate, that’s 5 to 10 percent. Nearly a tenth of all 200cm Japanese people end up absorbed into competitive sport.

Statistically, they’re too rare to observe. Athletically, they’re too rare not to be found.

The cause is the same, but the outcome flips. Because the population is thin, it falls below the resolution of any sampling design; and because the population is thin, the search cost drops low enough that everyone gets discovered. The same variable produces a failure of observation on one side and a success of observation on the other.

Trying the Netherlands (looks tall) as a test case

Look at the Netherlands as a control group, and this flips cleanly.

The mean height for adult Dutch men is 183.8cm, with an SD of around 7.1. 200cm is only 2.28σ. In frequency terms, that’s about 1 in 90 — the kind of rate where you’d find a few people per school year. Across 7 million adult men, that comes to 78,000 people.

And in the Netherlands, conscription-registration physical measurements cover nearly the entire male population of the relevant age, roughly 100,000 a year. The statistic exists.

Meanwhile, the number of pro-league slots in volleyball or basketball for a 200cm player is, at most, 100 to 200. The utilization rate is 0.2 percent. Of the 78,000 people, 99.8 percent are just ordinary office workers or teachers.

They’re counted, and nobody gets scouted. A perfect mirror image of Japan.

In Japan, being 200cm is, on its own, a variable that determines career choice with fairly high probability. In the Netherlands, it only gets you through the first filter, after which skill and motivation do all the work. The same attribute is close to a sufficient condition on one side, and merely a necessary condition on the other.

Even with similar averages, a different distribution shape changes the outcome

You might think this comes down simply to the Dutch average being higher.

But China tells a different story.

The average height of adult Chinese men, per the State Council’s figures, is 167.1cm. Calculated from provincial figures, it comes out to 169.7cm — lower than Japan. And yet China consistently fields 200cm-class players.

What matters isn’t the average, but the shape of the distribution.

By province: Shandong 175.44cm, Beijing 175.32cm, Liaoning and Heilongjiang 175.24cm. The lowest, Guangxi, comes in at 168.96cm. A 6.5cm gap between provinces. In Japan, prefecture-to-prefecture differences fall within statistical noise, and that’s the fundamental difference here.

A distribution formed by mixing groups with different means has fatter tails than a single-peaked distribution. And at the extreme tail, the higher-mean component ends up dominating almost entirely. This variance effect matters more than simply having 10 times the population.

It’s not the mean that builds the tail. It’s the spread.

That said, the numbers need a caveat. Hiroshi Mori of Senshu University has written that it isn’t even clear whether a reliable national average height exists for China at all. The provincial tables also come from private sources, with unclear survey years and age ranges. Read the figures cited here as order-of-magnitude sense only.

At 190cm, Japan is already on the Dutch side

There’s one more thing I want to verify with numbers.

At 190cm, the rate in Japan is also 1 in 1,400. High school boys across three grade levels number about 1.4 million, so nationwide there should be roughly 660 people at 190cm or above.

Look at the Japanese national football team: of the 26 players named for the 2026 World Cup, 2 are 190cm or above. The average was 181.5cm — ranked 36th out of 48 countries. Not tall at all.

But football only needs, at most, a few dozen players at the 190cm class, out of a pool of tens of thousands. The utilization rate is well under 1 percent. This is the same regime as the Netherlands’ 200cm case. Height only gets you through the first filter; whether you’re actually selected comes down to other variables.

So a 190cm player on the national team wasn’t chosen because he’s big. He’s winning on everything else.

The only height at which it becomes a sufficient condition in Japan is 200cm. And against an annual total of 350,000 male births, that 200cm mark shows up only 3 or 4 times per birth-year cohort.

Something too rare to count is nevertheless being counted down to the last one, and put to use.

References

This piece was conceived and directed by Kuzuryu, with the writing done by AI.

Originally published in Japanese at https://clazytech.com/2026/09/1763/. Translated with LLM assistance and reviewed before publication.