Final Prediction for the 2027 Boston Marathon Cutoff Time

It’s the week we’ve all been waiting for: registration for the 2027 Boston Marathon. Over the course of the past year, I’ve been tracking results from marathons across the country (and the world) to project the likely cutoff time. The last races were held last Sunday, and at this point data is all collected.

So it’s time to take a deeper look at how things changed from the last qualifying period to this one – and what means for a final prediction for the cutoff time for the 2027 Boston Marathon.

If you want the short version, the tracker is currently projecting a cutoff time of 5:13. But based on some other factors, I expect the actual cutoff time to be slightly higher. I’m pretty confident the cutoff time will be above 5:00, and it’ll likely be in the ballpark of 5:15 to 5:45.

Keep reading for a more thorough explanation of how we got here, how I came to this projection, and what is still in flux.

What Data Was Used For This Analysis?

Each week, I track the results from a sample of marathons across the United States (and the world). The main dataset includes every race in the United States and Canada with more than 200 finishers, the London and Berlin Marathons, and a handful of smaller races that are of particular significance.

To the extent that I can do so, I match multiple results from a single runner and focus on their fastest time. When each race is added to my database, I use fuzzy matching to compare the runner’s age, gender, and name to the existing results and look for duplicates.

Here’s an overall summary of the data:

Qualifying Period	2026	2027
Races	269	264
Finishers	624,149	686,891
Unique Finishers	559,612	623,457
Qualifiers	70,673	72,050
Unique Qualifiers	61,806	62,848

I also collected the results from the Tokyo and Sydney Marathons. However, due to the low conversion rates of qualifiers from those marathons, those results aren’t incorporated directly into the tracker. More on that later.

How Does the Boston Marathon Cutoff Time Tracker Work?

The tracker offers a simple projection of a potential cutoff time, based on a set of assumptions. You can read more about these assumptions and the calculations at the bottom of the page.

Assumption #1: The cutoff time is determined primarily by the number of applicants.

Assumption #2: A consistent percent of qualifiers will actually apply to run the Boston Marathon.

Assumption #3: The change in the number of qualifiers should be proportional to the change in the number of applicants.

Assumption #4: There’s a stable, linear relationship between the percentage of applicants cut and the cutoff time.

Therefore, we can a) estimate the number of applicants, b) estimate the number of applicants cut, and c) estimate the cutoff time.

Last year, there were 33,249 applicants for the 2026 Boston Marathon. The number of unique qualifiers increased 1.69%, which would suggest an expected total of 33,810 applicants. If 24,362 applicants are accepted, that would yield 9,448 rejections. That 28% rejection rate, fed back into the linear regression equation yields a projected cutoff time of 5:13.

What Factors Influence the Number of Qualifiers?

There are a few things that influence how many qualifiers were produced throughout the qualifying period.

First, the qualifying times stayed the same this year. Last year, when the qualifying times were reduced, this caused a significant drop in the number of qualifiers between the 2025 and 2026 qualifying periods. That big drop will not occur this year.

Second, there was a minor change to the qualifying process. BAA implemented a new downhill results index, penalizing races with excessive net elevation loss. Based on this analysis, the rules themselves directly eliminated a small number of qualifiers (383). However, indirectly these changes caused a large decline in finishers at some races and the impacted races produced ~1,800 fewer qualifiers this year. Some of that decline was likely offset by increases elsewhere.

Which brings me to the third change – the net increase in finishers. Across all races, there was an increase of ~60,000 finishers (10%). If you exclude the Majors, which have less room for growth, there was a net increase of 15.6% across all other races. All things being equal, you’d expect more finishers to also result in more qualifiers.

The final factor is race day weather. At a high level, this should wash out as weather fluctuates from one weekend to another. But bad weather at a particular race or on a particularly busy race weekend can have an outsize influence. This is the case for Berlin, where it was very hot. But there were a number of other races weekends with warm weather, dampening the potential growth in qualifiers. On the flip side, Boston had excellent weather and the number of qualifiers from that race increased dramatically.

Based just on the growth in finishers, you’d expect to see closer to a 10% increase in qualifiers. That would have pointed towards far more applicants (~36,500) and a much higher cutoff time (~6:20). The weather at Berlin and other races instead kept the number of qualifiers – and the cutoff time – at a more reasonable level. But this is, to some extent, luck. And it’s possible that better weather next year will yield a much larger qualifying field.

Where Are the Potential Errors In This Methodology?

This is not perfect, and there are a few things that could influence the outcome one way or another.

The first is the results from Berlin. This methodology essentially treats all races the same, but some races see a smaller share of their qualifiers convert to actual applicants. Berlin is one of these races. There was a huge drop in qualifiers at Berlin due to the weather, and this held the projection down all year. If you set Berlin aside, there would have been an 8.74% increase in BQ’s. That would point to a 6:11 cutoff time. Obviously, you can’t totally ignore Berlin – but if you split the difference you get a cutoff time of 5:30 to 5:45.

The second is the results from Tokyo and Sydney. These aren’t factored into the tracker at all because the conversion rate of those BQ’s is much lower. Year over year, these two races did see an increase (~600 at Tokyo and ~100 at Sydney). However, there were two Sydney Marathons (2024 and 2025) in the last qualifying period. So there’s really a net decrease of ~1,600. If we treated these like other races, that would wipe out the actual net increase in BQ’s and suggest a cutoff time of just under 5:00. The impact is likely much smaller, but if you split the difference this could nudge things down 5 seconds or so.

A third issue is the actual conversion rate of qualifiers to applicants. This methodology assumes that it’s stable from year to year. Last year, it wasn’t the same as the previous year. Perhaps due to the change in the qualifying times, the percent of runners with certain buffers that applied was different from the 2025 qualifying period. Since there was no change to the qualifying times, I’ll assume that this won’t be an issue this year.

A final issue is the actual number of accepted applicants. Last year, that was 24,362. That’s the highest it’s been in recent years. It’s slowly crept up over the past few years. But this year they are adding 1,000 spots for a lottery. The math here assumes that the number of accepted applicants remains 24,362. With the lottery, I highly doubt it will increase further. But it could decrease slightly. Taking the number of applicants back down to 24,000 – a nominal decrease – would bump the projection up almost 15 seconds.

What About the Distribution of Buffers?

One more issue to consider is whether this year’s qualifiers are similar to last year’s qualifiers. And they aren’t identical.

Specifically, if you look at the distribution of buffers, this year’s qualifiers are faster. Overall, there’s a small net increase. But there’s a much larger net increase across runners who have a 10+ minute buffer. This is offset by a decrease among runners with a buffer under five minutes. Since the fastest runners are accepted first, this could mean that the field would actually fill up faster.

A different way to look at this is to use the conversion rates from last year’s race to calculate how many applicants we should expect across the full spectrum of buffers. By graphing these two lines, along with information from last year’s cutoff, we can project what cutoff is to be expected from this qualifying pool.

In the visual above, the red line is the cumulative number of applicants across the distribution last year. The green line is this year. Notice that they nearly intersect at the end – because there are only slightly more qualifiers this year than last year. But there’s a bigger gap in the middle, pointing to the large increase in qualifiers with 10 to 20 minute buffers.

The dotted lines indicate three things.

The vertical line on the right marks out last year’s buffer: 4:34.

The horizontal line marks out the calculated number of applicants (24,239) at that cutoff.

The vertical line on the left is the approximate point where that horizontal line meets the green curve. This should indicate the likely cutoff time: approximately 5:09.

I would expect this estimate to be more accurate than the more general one used by the tracker. After all, that overestimated things last year, so if we’re looking at the net change from last year’s projection we could once again be overestimating things.

But this is pegging things against the actual cutoff from last year – 4:34 – and taking into account the distribution of buffers among the qualifiers and expected applicants.

What’s the Bottom Line?

Up until a month or two ago, I thought that a worst case scenario above 6:00 was still possible. The last few weeks have been kind to those who have been patiently waiting to see if their buffers would be good enough – although those weeks haven’t been kind to many of the people trying to notch a last minute qualifying time. At this point, that worst case scenario is out the window.

The basic math of the Boston Marathon Cutoff Time Tracker points towards an expected cutoff of 5:13, although that regression formula could be off by 20 or 30 seconds in either direction.

But a more detailed analysis of the data, utilizing the actual buffers that individual runners have achieved, points towards a cutoff of 5:09.

Next take into account the fact that a) the Berlin results are weighting this down and b) the number of time qualifying bibs is unlikely to increase (and may decrease slightly) and the odds of this being higher than 5:09 is much greater than the odds of it being lower.

There’s an inevitable amount of variation in the actual outcome, so it’s anyone’s guess where it actually lands. But when I combine this data, these factors, and my gut, I’m going to guess the actual cutoff is in the range of 5:20 to 5:30. I’d be very surprised if it was below 5:00 and I’d be equally surprised if it was above 5:45.

Based on previous years, we can expect to learn the number of actual registrants early this week. However, this number will likely be influenced by the existence of the lottery. So it may be difficult to parse exactly what it means for the cutoff time. It’ll likely be significantly higher than last year, but it’s unlikely that it’ll hit record levels. I’d expect that number to be 35,000 to 36,000.

In any case, the die is cast. Good luck to everyone who registered – whether you’re waiting for the cutoff announcement or the lottery.