Last verified against the paper behind the weekly mileage model (Vickers and Vertosick, 2016), the runner data its authors published, and the B.A.A. qualifying page.
Example: marathon from a 50:00 10K and 20 miles a week
The weekly mileage model, built from a survey of 2,303 recreational runners, predicts 4:11:51 for a 50:00 10K on 20 miles a week. Riegel's formula uses the 10K alone and says 3:50:01, 21:50 faster. Enter your own race above to replace this example.
| Other prediction | Time | Per mile |
|---|---|---|
| RiegelThe 10K alone | 3:50:01 | 8:46 |
| CameronThe 10K alone | 3:54:16 | 8:56 |
The weekly mileage model (Vickers and Vertosick, 2016, model 1), speeds in meters per second: speed = 0.16018617 + 0.83076202 × your Riegel speed at the authors' exponent of 1.07 (the Riegel row above uses 1.06) + 0.06423826 × weekly miles ÷ 10. It assumes an average course.
Example: men, age 40, 3:04:00
A buffer of 1:00 would have got a place in none of the last 4 Boston fields (the 2024 to 2027 races). Each year the B.A.A. sets a cut-off: how far under the standard a qualifier had to be to get a place. Some of those cut-offs were set against a different standard for your age group, so read them as a rough guide.
| Race | Cut-off | Your buffer against it |
|---|---|---|
| 2024 race | 5:29 | 4:29 short |
| 2025 race | 6:51 | 5:51 short |
| 2026 race | 4:34 | 3:34 short |
| 2027 race | 5:17 | 4:17 short |
Buffer = your 2027 qualifying standard minus your chip time. Standards and cut-offs from the B.A.A. qualifying page, checked 4 Oct 2026.
How the three predictions work
Riegel multiplies your time by the ratio of the two distances raised to the power 1.06. It predicts the 10K and the half marathon well. For the marathon it tends to come out fast. Riegel's 1981 paper gives 1.08 for world-class runners and 1.05 to 1.06 for men aged 40 to 70. We use 1.06.
Cameron scales your time along a speed curve that slows as the distance grows. Like Riegel, it uses one race and nothing about your training.
The weekly mileage model comes from a 2014 survey of 2,303 recreational runners. It starts from your Riegel speed, worked out with an exponent of 1.07 (its authors' choice), and adds your weekly mileage. In the survey, plain Riegel predicted the marathon 10 or more minutes too fast for about half the runners, and the weekly mileage model did so for about a quarter.
The weekly mileage model takes one race: a 5K, 5 miles, 10K, 10 miles or a half marathon. It assumes an average course, and the runners it learned from ran 10 to 103 miles a week. Outside those limits, read the time as a rough guide.
Boston Marathon qualifying standards for 2027
The Boston qualifier mode above takes your age on Boston race day and your marathon chip time, finds your qualifying standard, and shows how far under it you are. That buffer is set against the cut-offs of the last four races: each year the B.A.A. accepts qualifiers from the fastest down until the field is full.
The B.A.A.'s qualifying times for the 2027 race, by age on race day. Non-binary runners have the same standards as women. The B.A.A. has not published the 2028 standards yet (checked 4 Oct 2026).
| Age on race day | Men | Women | Non-binary |
|---|---|---|---|
| 18 to 34 | 2:55:00 | 3:25:00 | 3:25:00 |
| 35 to 39 | 3:00:00 | 3:30:00 | 3:30:00 |
| 40 to 44 | 3:05:00 | 3:35:00 | 3:35:00 |
| 45 to 49 | 3:15:00 | 3:45:00 | 3:45:00 |
| 50 to 54 | 3:20:00 | 3:50:00 | 3:50:00 |
| 55 to 59 | 3:30:00 | 4:00:00 | 4:00:00 |
| 60 to 64 | 3:50:00 | 4:20:00 | 4:20:00 |
| 65 to 69 | 4:05:00 | 4:35:00 | 4:35:00 |
| 70 to 74 | 4:20:00 | 4:50:00 | 4:50:00 |
| 75 to 79 | 4:35:00 | 5:05:00 | 5:05:00 |
| 80 and older | 4:50:00 | 5:20:00 | 5:20:00 |
Boston Marathon cut-offs since 2012
How far under the standard a qualifier had to be to get a place in each race, and how many qualifiers missed out. A cut-off of 0:00 means every qualifier who applied got in. The B.A.A.'s table lists no cut-off for the 2013 race, so neither does this one.
| Race | Cut-off | Qualifiers not accepted |
|---|---|---|
| 2027 | 5:17 | 8,019 |
| 2026 | 4:34 | 8,887 |
| 2025 | 6:51 | 12,324 |
| 2024 | 5:29 | 11,039 |
| 2023 | 0:00 | 0 |
| 2022 | 0:00 | 0 |
| 2021 | 7:47 | 9,215 |
| 2020 | 1:39 | 3,161 |
| 2019 | 4:52 | 7,248 |
| 2018 | 3:23 | 5,062 |
| 2017 | 2:09 | 2,957 |
| 2016 | 2:28 | 4,562 |
| 2015 | 1:02 | 1,947 |
| 2014 | 1:38 | 2,976 |
| 2012 | 1:14 | 3,228 |
Race prediction questions
Which prediction should I trust for a marathon?
The weekly mileage model, if you entered your mileage. In the study, its predictions came closer to real marathon times than plain Riegel's.
Why is my marathon prediction slower than Riegel's?
Riegel's formula knows your race time and nothing else. The weekly mileage model also counts your training, and in its survey, runners on fewer miles a week slowed more over the marathon than Riegel expects.
Can I predict a 5K from a marathon?
Yes. Pick Marathon as your recent race. Riegel and Cameron work in both directions; the weekly mileage model predicts the marathon only.
Does the Boston qualifier tell me if I will get in?
No. It shows your buffer and how that buffer compares with the cut-offs of the last four races. The cut-off depends on how many qualifiers apply each year, so no buffer is certain in advance.
Sources
- Vickers AJ, Vertosick EA. An empirical study of race times in recreational endurance runnersBMC Sports Science, Medicine and Rehabilitation 2016;8:26. The weekly mileage model is their model 1: coefficients from Additional file 1, the authors' own predictions in Additional file 2. CC BY 4.0. Read 4 Oct 2026.
- Riegel PS. Athletic Records and Human EnduranceAmerican Scientist 1981;69(3):285-290. Page 286 gives the exponents: 1.08 for world-class runners, 1.05 to 1.06 for men aged 40 to 70. Read 5 Oct 2026.
- D. F. Cameron's race time model, published by P. HoffmanArchived copy (10 May 2013) of the University of Massachusetts Lowell page, with Cameron's formula and his own limits.
- B.A.A. Boston Marathon qualifying standardsBoston Athletic Association. The 2027 standards and the cut-off for every race since 2012. Read 4 Oct 2026.
- World Athletics Technical Rules (Book C, C2.1)In force from 1 July 2026. The race distances.
- NIST Special Publication 811, Appendix B.8NIST, 2008 edition. 1 mile = 1.609344 km, an exact factor.
Spotted something out of date? Report an error.
