The marathon record is faster than every Boston finisher
I took the whole 2014 Boston Marathon field, fitted the distribution its finish times form, and marked Sabastian Sawe's 1:59:30 world record on the same axis. The race winner ran 2:09. The median runner took 3:52, which is 94% longer than the record.
78% of finishers ran between 3:00 and 5:00. Gray marks the rest of the field. Nobody in it ran inside the gap between the record and 2:09.
- 2.4 SD
- The record sits 2.4 standard deviations below the field mean of 4:03 (SD 51 minutes).
- Log-normal
- A shifted log-normal fits the body of the field: right-skewed, peaking around 3:30 to 3:50, with a long slow tail.
- ξ = −0.15
- A generalized Pareto fit (Peaks-Over-Threshold) on the fastest 5% gives a negative shape parameter, so the fast tail is bounded.
The field, with the record marked
Fitted, the field is log-normal, with a median of 3:52 and a slow tail that pulls the mean to 4:03. Sawe's record lies to the left of the whole curve, winner included.
Nobody in the field ran inside the gap between 1:59:30 and 2:09
Hover the chart, or focus it and use the arrow keys, to read a bar.
tables/web_continuum.json, written by
export_web.py from the file analyze.py reads.Look up a finish time
h:mm:ss or h:mm, between 2:00 and 7:00
The fixed rows from tables/you_vs_goat.csv: the share of this field a given time finishes ahead of, and how much slower it is than the record.
| Finish time | Ahead of this share of the field | Slower than 1:59:30 by |
|---|---|---|
| 2:30 | 99.7% | 25.5% |
| 2:45 | 98.1% | 38.1% |
| 3:00 | 92.6% | 50.6% |
| 3:30 | 70.6% | 75.7% |
| 4:00 | 43.2% | 100.8% |
| 4:30 | 26.3% | 125.9% |
| 5:00 | 14.4% | 151.0% |
| 6:00 | 2.8% | 201.3% |
The fast edge has a limit
The body fit says little about the extreme fast end, so the tail gets its own model. I took the fastest 5% of finishers (1,593 runners under 2:56) and fitted a generalized Pareto distribution to how far each one beat that threshold. The shape parameter came out at ξ = −0.147, and a negative shape means the tail ends at a finite point. Because the method is unreliable at locating that point, I don't quote a fastest possible time.
What this doesn't show
Boston is a qualifying race, so everyone in this field was already fast enough to enter. Every percentile on this page is about this one field; a general-entry marathon would sit slower and push the record further out.
The record on this page is Sabastian Sawe's 1:59:30 (London, 26 April 2026), which World Athletics ratified on 23 July 2026; it is the value hard-coded in analyze.py. It replaced Kelvin Kiptum's 2:00:35, and the change moved the record mark and the percent-slower figures and none of the field statistics.
I don't quote the record's rarity as "1 in N". That would mean trusting the fitted density far beyond the data, so the record is described by what was observed: faster than all 31,842 finishers, 2.4 standard deviations from the mean.