Boston Cutoff Time: What the Number of Applicants Tells Us About a Prediction

Friday, the registration period for the 2025 Boston Marathon came to a close.

Along with many others, I was eagerly waiting to see if they would release the number of qualified applicants. They didn’t.

By the end of the weekend, I was beginning to think they might withhold that information until they announce the actual cutoff time. But today, that changed.

The BAA officially announced that there were 36,406 qualified applicants for the 2025 Boston Marathon. This is a record number – up about 10% from last year’s 33,058.

So what does this mean for a possible cutoff time?

The Original Prediction and The Unknowns

When I initially analyzed the data, I came to a prediction of 7:03 for a cutoff time.

I followed that up with some further analysis, exploring a few factors that could influence the final outcome.

The balance of the evidence suggested that a 7+ minute cutoff time was likely. But there were two unknowns that could mitigate that: the Easter effect and the field size.

Some people argued that the Easter holiday would discourage people from registering for the race. If that were true, it would depress the number of applicants and reduce the cut-off time.

Some people also held out hope that BAA would increase the field size – either adding additional seats (like they did in 2020) or reserving more seats for time qualifiers. If either of these things happened, it would reduce the need to eliminate some qualified applicants – and marginally reduce the cutoff time.

So what can we learn from today’s announcement?

Easter Had No Effect

I think one thing should be clear: Easter had little to no effect on the number of applications.

My prediction for the number of applicants – without taking Easter into account – was 36,248. This was based on an analysis of the number of qualifiers in my large dataset combined with the likelihood that a given runner would apply based on their BQ buffer.

That’s pretty close to the actual number – 36,406.

There was one year in which Easter directly preceded the race and in which the number of applicants seemed abnormally low – 2017. If Easter explained that dip, then Easter could account for about an 8% reduction in the number of applicants.

So if Easter did have a large effect, the total number of applicants would likely have been in the 33,000 to 34,000 range.

The fact that it exceeded 36,000 and lined up with a prediction based on a model that did not account for Easter at all tells me that the effect of Easter is likely negligible.

The Field Size Will Remain 30,000

In its announcement, the BAA announced that there would be 30,000 runners at this year’s Boston Marathon.

That strikes another potential mitigating factor from the list of possibilities. The overall field size will not be increasing.

In this article, I reviewed the history of the field size and the number of qualified applicants who were accepted. There was one year – 2020 – in which they pushed the field size to 31,500 to accommodate a surge in applicants. That isn’t going to happen again.

The final unknown, however, is how many applicants will be accepted. Last year, this dropped from ~23,000 to ~22,000. BAA could limit some of the other forms of entry and restore those 1,000 seats.

That would marginally impact the cut-off time. But those 1,000 seats would only push the cut-off time down by around 30 to 45 seconds.

Has the Prediction Changed?

Based on this evidence, my confidence in the original prediction is stronger today than it was last week.

The model is based on understanding a) how many people qualify for the Boston Marathon, b) how big individual runners’ BQ buffers are, and c) how likely runners are to apply based on their individual BQ buffer.

That got me pretty close to the actual number of applicants. And with that confirmation, things are becoming increasingly clear.

The visual above is a breakdown of the number of predicted applicants based on how big their buffer is – compared to the estimated number of applicants per minute of buffer in 2024.

If the number of spots remains 22,000, the field will be full right around the 7 minute mark.

• 7,050 applicants will have a 20+ minute buffer
• 9,710 applicants will have a 10-20 minute buffer
• 4,960 applicants will have a 7-10 minute buffer

That brings the total to 21,720. An additional 1,920 applicants will have a 6-7 minute buffer.

So the applicant pool should be full right around the 7 minute mark – and I’d feel comfortable narrowing the margin of error down to 6:30 to 7:30.

Are There Still Any Unknowns?

There are really only two things left that we don’t know.

First, the number of actual available seats. This prediction assumes the size of the qualifying field remains the same at 22,000. If it increases by 1,000, then that would reduce the cut-off score by about 30 seconds.

Second, whether the distribution of buffers is predictable based on last year. It’s possible that fewer people with 0-5 minute buffers applied, because they assumed they wouldn’t get in. This would in turn mean that there are relatively more applicants with big buffers. If there are more applicants with 10+ minute buffers, that would increase the cut-off time.

But I don’t think this is likely. It’s not enough that the 0-5’s didn’t apply. This scenario would also require that more of the 10+ runners applied. Otherwise, the math would work out to get to the same total number of applicants. And there’s no reason to think that runners with 10+ minute BQ buffers would act any differently.

Nonetheless, I’d say the worst case scenario is around 8:00 to 8:30. But if I was betting money, I’d cap the likely outcomes at 7:30.

Now, we wait.

Based on last year, it’ll be the end of next week when the cut-off time is announced. It’s possible the pre-verification process could speed that up, but there’s no indication from BAA’s announcement today that that’s the case.