Tokyo Marathon Part 3: How Many Runners Qualified for the Boston Marathon?

Feature photo by Naoki Nakashi on Flickr – CC BYSA 2.0

It’s been a couple of weeks now since the Tokyo Marathon. I wrote up a general analysis of the data from the race, and I followed that up with an exploration of how the field varied throughout the corrals.

But that preliminary data was missing a key component: age data.

That made it impossible to really make apples to apples comparisons of how the field compared to last year. It also prevented me from exploring a question that many people are interested in – whether this will have any impact on the Boston qualifying cutoff time.

They have since released the official results, and I collected the age data to go with the other information that I have.

Two topline stats:

• The number of people meeting their BQ dropped from 6,064 in 2024 (using the old qualifying times) to 4,801 in 2025 (using the new qualifying times. This is ~21%, which is more or less what you’d expect from a race with a similar field size.
• Applying the new qualifying times to both years, the percent of runners meeting their BQ dropped from 14.5% to 13.3%. This suggests either a) the composition of the field shifted, b) the heat took a minor toll, or c) a combination of the two.

With that in mind, let’s dive in and dig a little deeper.

Change in the Number of Qualifiers by Age Group

Looking at the field as a whole can mask a lot of smaller changes happening throughout the field. The qualifying times effect different age groups differently, and it’s possible that the actual composition of the field has shifted.

So what if we take the absolute numbers – how many people qualified – and break that out by age and gender.

In the visual below, the top chart represents men and the bottom chart represents women. The left bar represents the number of runners in 2024 who met their old BQ, and the right bar represents the number of runners in 2025 who met their new BQ.

The biggest differences happen with men 40-59 and women 45-54. The men saw 25-30% drops throughout that range, while women was a 20% drop in the 45-49 age group and a 30% drop in the 50-54 age group.

Notably, there are some minor decreases to the right – among runners 60+. This can’t be attributed to the new qualifying times, since those runners still have the same qualifying times.

Meanwhile, among women the number of 25-29 year olds qualifying stayed the same. And among both men and women, the dropoff among younger runners was much smaller than you’d expect.

The Tokyo Marathon trends much older than the other Majors, and this is likely a sign that the age distribution shifted to the left between last year and this year.

How Many Qualifiers Are Likely to Apply?

Another really important piece of the BQ puzzle, though, is whether these runners are actually going to apply to run.

Although 4,800 qualifiers is a big drop from 6,000, it still represents a lot of runners who met their BQs. I haven’t incorporated this data directly into the Boston Marathon cutoff time tracker, and some people have asked me why.

My typical response is that – of all the Majors – Tokyo is the least likely to have runners convert to Boston applicants. Based on BAA data, we know that Chicago, London, and Berlin are among the top five races.

Tokyo was not on that list. And after analyzing the field at the 2024 Boston Marathon and the entry list for the 2025 Boston Marathon, Japanese runners are far down on the list of the most common countries at Boston. And Japanese runners still make up a majority of the field at Tokyo – even though that has been dropping in recent years.

To illustrate how important this is for understanding the BQ numbers, let’s segment the race into four groups – Japanese runners, American runners, the next five most common countries at Boston, and everyone else. The number of BQs are graphed in the visual below, broken out by men (top) and women (bottom) with last year’s number in the left bar and this year’s number in the right bar.

Note: the top five countries represented here are Canada (1,799), Great Britain (1,384), Mexico (719), Brazil (587), and China (505). The numbers in parentheses are the number of runners from that country in the 2025 Boston Marathon entry list – compared to 21,412 runners from the United States and 195 from Japan.

The leftmost set of bars represents runners from Japan, and they make up the majority of Boston qualifers at Tokyo. Tokyo is, notoriously, weighted heavily towards men, and that’s especially true of the Japanese part of the field. So the difference is more stark when you look at the men.

Collectively, Japanese men and women represent 2,697 of the BQs are this year’s Tokyo Marathon.

All the way to the right, you see the bars for the United States. The number of American runners is on the rise, so that contributes to the smaller difference between 2024 and 2025. But still, there are only 574 BQs from the US in 2025.

The other top countries at Boston represent 570 of the qualifiers. So collectively, the top six countries represented at Boston account for less than 25% of the qualifiers at this year’s Tokyo Marathon.

Incorporating all of these qualifiers into the cutoff tracker would add a lot of irrelevant data and likely distort the outcome. The most relevant question is – how did the pool of likely applicants change?

And among those two groups, the number of qualifiers dropped from 1,362 to 986 – a 16% drop. That’s a larger drop than the current year to date drop across the entire field (~8%), but it’s also not a very large one in the scheme of things.

If you were to ask me what this means for the cutoff time, I’d say the results at Tokyo are a wildcard – which could indicate a slight moderation of the overall projection.

Did the Age Distribution of the Race Change?

As I mentioned earlier, the Tokyo Marathon trends older compared to other races. In part, this is due to the age distribution of the Japanese population – and of Japanese runners.

But over the past few years, the share of Japanese finishers at Tokyo has been trending downwards and the share of American runners has been trending upwards. This has shifted the gender distribution slightly, but is it also shifting the age distribution?

The visual below is divided into men (top) and women (bottom) and compares last year (left bar) to this year (right bar). Each group of bars represents an age group – from under 20 to 75-79.

The general distribution is the same – with the 50-54 age group being the largest, followed by the 45-49 age group. But within the existing distribution, there is definitely a shift.

Among men, every age group below 40 grew. The largest change, among men 30-34, amounted to a 34% increase.

Among women, every age group under 45 grew. The largest increase, also among women 30-34, amounted to a 26% increase.

Given the fact that the overall field size is roughly the same, if one area grows that must mean another area is shrinking. The older women don’t see much change, but there are pretty significant decreases among the older men.

Every age group from 45-49 on sees a decrease, with the biggest concentration among men 50-59.

When you put it altogether, older runners – who are typically older, Japanese men – are being replaced by younger runners – who are increasingly international and a relatively more even mix of men and women.

Did the Boston Qualification Rate Change From 2024-25?

Earlier, we saw that the absolute number of Boston qualifiers did indeed drop significantly from 2024 to 2025. And, even if you applied the same standards to both years, there would still be a small drop.

But there was also a pretty significant shift in age from year to year, which could confound with the number of runners who qualify year to year.

So for one final look, let’s apply the same standard (the 2026 BQ times) to both this year and last year and visualize the qualification rate in each age group.

The largest age groups are men 40-59. And among these age groups, there’s a slight decline in the qualification rate. It drops by one percentage point in a couple of areas, and less than one percent in others.

Among the largest groups of women, too, there isn’t much difference year to year.

The biggest decreases occur among younger runners – 25-39.

At the other end of the spectrum, there’s a jump in qualification rate among runners 70-79. Although it’s worth pointing out that these are very small groups, and this could just be the presence of a handful of outliers.

So what should we make of this?

On the one hand, I think this confirms the general conclusion that I arrived to in the first analysis – that the heat didn’t play a huge role in the results. If it did, you’d see a consistent drop across all of the age groups. The stability – especially among the largest age groups – suggest that environmental factors didn’t have a huge impact.

On the other hand, there was a decrease in the qualification rate among the younger runners. This is also where the largest increase in the number of finishers occurred, so it suggests that the new runners coming in may not be as fast – on average – as the existing pool.

Also, it’s worth pointing out – again – that the qualification rate among runners 60+ is much higher than it is for the other age groups.

Another Look at the Distribution of Finish Times

One of the more controversial conclusions in my analysis was that the heat didn’t have a big impact on the overall distribution of finish times. This was based on looking at the overall distribution – and there was an inherent risk there that shifts in the demographics of runners could have masked differences in times.

To make an apples to apples comparisons, you want to identify similar age groups and see if the distribution of finish times changed within them.

In the two visuals below, I’ve isolated the 20-44 age groups for women (first) and men (second). In each visual, the overall distribution of finish times – segmented into five minute bins – is graphed for 2024 and 2025.

If you hover over a dot, you’ll also see the cumulative percent of runners who beat that time. Note that through a quirk of the graphing software, the cumulative number is not formatted – so 0.07 means 7%.

In this visual, you see the overall distribution of finish times for young women. And when you compare the two lines, there isn’t much difference at all.

The 2025 line is a tiny bit lower from 3:00 to 3:30, and it’s significantly higher at 3:40 and at a couple spots. But if you compare the cumulative levels at any given point, they’re pretty similar. Maybe off by a percentage point here and there, but there’s no consistent difference.

Here’s the visual for the men, and the overall trend is again similar. Among the fastest men, last year was slightly better. But throughout the bulk of the distribution, there’s no clear difference where one year is higher than the other.

If heat played a big role, you’d expect everyone to have lost time – so that the overall distribution of finish times was pushed a couple of minutes to the right. And this difference would grow as you got into the slower parts of the field.

And maybe if you segmented this down to the minute – instead of five minute buckets – you’d see a small shift. But if anything, this supports the conclusions that the heat had little or no effect on the finish times. There’s just simply no evidence here – when you look at all of the finish times – that there was a widespread effect.

So What Did We Learn About the Tokyo Marathon?

If we step back and combine these three analyses, here are the general conclusions that we’ve reached.

First, the demographics are shifting.

Historically, this race has been dominated by Japanese runners and these are overwhelmingly male and old(er). As the international share of runners increases – especially Americans – this is shifting the gender distribution towards women and the age distribution towards younger runners.

Second, the heat did not have a broad effect in finish times.

This isn’t meant to negate any individual runner’s experience. Certainly, an individual could have suffered from the heat. But when you’re trying to understand what happened at scale, you have to zoom out. And no matter how we looked at the data, there’s simply no big difference in finish times from 2024 to 2025. If there was a weather effect, it was a minor one.

Third, the race isn’t likely to have a huge effect on the Boston Marathon cutoff time – but it is a factor that should be considered in the final analysis.

While the absolute number of qualifiers dropped significantly (~21%), the majority of runners come from countries that don’t typically run Boston. When you narrow in on the countries likely to run Boston, the drop is smaller (~16%) and the number of runners in question is much smaller.