Update to “An attempt to predict the cutoff time for the 2025 Boston Marathon.”
In the nearly three years that I have been writing this blog, each of my posts have received at best about 900 views.
Then, in December 2023, I made a prediction of the cutoff time for the 2025 Boston Marathon. To date, that post has received more than 11,000 views — far beyond the readership of my previous posts. Many readers are eagerly awaiting this new post wherein I provide an update to the earlier prediction.
(By the way, there is some good stuff in my earlier view-challenged posts. You may want to check them out. Just saying.)
This is new territory for me. I understand that some runners may modify their strategies based on what I write. I feel an obligation to be as thoughtful and as transparent as possible in providing this update.
Hence, before jumping in with my update, I’ll clarify the thinking regarding my method for arriving at a prediction. Readers’ comments on the earlier post suggested a need for the clarification.
As I pointed out in the December post, any attempt to predict the cutoff time starts with an estimate of the number of qualified runners (BQ’s) who will apply to run in the 2025 Boston Marathon. Given the fixed field of 30,000 entrants to Boston, the more applicants the larger the cutoff time.
Every year there are far more BQ’s achieved worldwide than there are Boston Marathon applicants. Why the difference? I can only speculate: runners with more than one BQ in a single qualifying period, runners who wanted to run in Boston but encountered a precluding conflict, fast runners who, remarkably, never had any intention of running in Boston, etc.
How does one pare down this list of BQ’s to arrive at a prediction for the number of applicants?
My approach was to look for correlations and I found one between the number of applicants and the total number of BQ’s achieved from the top 50 North American (NA) qualifying races, that is, the 50 marathons that rank highest in the number of BQ’s generated. Instinct suggests that many entrants chose these races with the sole intent of applying to Boston.
I found a strong correlation between this sample set and the number of applicants to Boston for the 2018, 2019, 2023, and 2024 races (I am excluding the pandemic-disrupted years of 2020, 2021, and 2022 because they aren’t analagous to the 2025 race.) The number of applicants to Boston fell into a range of 78.1% to 80.2% (call this the reduction factor) of the total number of BQ’s generated by the top 50 NA qualifying races.
If there were a better sampling approach to arriving at the number of applicants to Boston, I would happily use it.
One reader pointed out that my method does not account for international runners. That’s fair. In response, it is worth referring to Brian Rock’s excellent summary of the numbers from this year’s Boston Marathon and the raw data supplied by the BAA.
Of the 29,333 runners entered in the 2024 Boston Marathon, 20,051 reside in the United States. The largest contingent of international entries comes from Canada (1705) followed by the United Kingdom (1236).
An overwhelming majority of the entrants were from North America (21,756 of 29,333) so I feel justified in heavily weighting my sample set towards North America, which includes Canada.
One reader (Geogeec), brought to my attention that the qualification period for the 2024 race included two London Marathons (2023 and 2022 because the 2022 race was held in October) whereas the 2025 qualifying period only holds the 2024 London Marathon.
It’s an important observation suggesting that the number of UK, and by association international, runners may decrease in 2025 as a result. However, my guess is that the difference will not be significant enough to change my sampling approach. I state this because in 2023, which included only one London Marathon in the qualifying period, there were 890 UK entrants, only 346 fewer than the double-London dose of 1236 for the 2024 race.
I mentioned in my December post that 29 of the likely top 50 NA qualifiers had not yet occurred with the most significant omission being Boston itself. That is why I was waiting for last week’s Boston Marathon before providing this update.
In 2023, the Boston Marathon with ideal weather conditions generated 13,741 BQ’s. This year, in oppressive heat, only 9,814 runners BQ’d. Consequently, applying the algorithm from my earlier post, my high end estimate for cutoff time drops considerably.
There has been some minor shuffling of the top 50 NA qualifyers for the 2024 Boston Marathon compared to the 2025 Boston Marathon qualifying period. By my estimate, 34 of the top 50 NA 2025 Boston Marathon qualifiers have taken place and have yielded (according to marathonguide.com) 39,536 BQ’s compared to 36,564 BQ’s at the same point in the previous qualification period, an 8% increase.
This contrasts sharply with the 30% increase in BQ’s I reported in December. Interestingly, the 30% increase was holding up until Boston last week but the 29% year over year decline in Boston BQ’s deflated that trend.
As I did in December, I choose to bound my estimate for the remaining 16 qualifying races, the most significant of which is Grandma’s Marathon in June. At the low end, the remaining races will yield the same number of BQ’s as the previous year’s race. At the high end, those races will return to the super shoe-powered 30% increase in BQ’s.
My December model multiplied my estimate for BQ’s by a 78.15% reduction factor to arrive at the number of applicants for the 22,019 available bibs. Both the 78.15% and 22,019 data points came directly from the numbers associated with the 2024 Boston marathon.
I recognize that during the reference years 2018, 2019, 2023, and 2024, the reduction factor varied from 78.15% to 80.20% and the number of available bibs ranged from 22,019 to 23,267. I take into account these spreads in my updated cutoff time prediction as shown in the table below. I am fitting these data to the same normal distribution parameters that I described in my December estimate in order to extract cutoff time estimates.
My December estimate ranged from 5:52 at the low end to 8:43 at the high end. The updated prediction for the 2025 Boston Marathon cutoff time is between 5:43 and 7:11.
The more recent data suggests that the 8:43 cutoff at the high end is no longer supportable by the results of the qualifiers that have taken place since December. Although many of the qualifiers have seen strong percentage increases in year over year BQ rate, most of them have too few entrants to make a big difference in the total number of applicants.
As always, I welcome comments and suggestions for improvements to the model.
