Data story · 14 years of Boston Marathons

The Heat Tax

Every runner knows heat ruins a race. I wanted the price tag — so I paired fourteen Boston Marathons with their race-day weather.

By Jeremy Lee

r = 0.86

correlation between temperature and field median time

~1 min

slower per extra degree Fahrenheit

+31 min

median cost of a hot 80°F vs cool 50°F day

The question

We all know a hot day makes a marathon harder. But "harder" isn't a number. By how many minutes does a degree of temperature actually slow a whole field of runners? Boston is a perfect natural experiment: the same iconic course, run every April, but under wildly different weather — from a chilly 45°F to a brutal 81°F.

What I did

I took the official results of every Boston Marathon from 2001 to 2014 and computed each year's median finish time — the time of the runner smack in the middle of the field. Then I pulled the race-day temperature for each year from a historical weather archive and put the two side by side.

Hotter years are clearly slower

The relationship is strong and hard to miss: the warmer the race, the slower the field, with a correlation of r = 0.86. The fit works out to roughly one extra minute on the median finish for every degree Fahrenheit. The two hottest years on record here — 2012 (81°F) and 2004 (74°F) — sit far above everyone else.

Scatter of race-day temperature vs field median finish time with upward trend — Each dot is one Boston Marathon. Hotter races (right) had slower fields (up). 2013 is shown hollow — its race was halted, so it is excluded from the fit.

Half an hour, gone

Stretch that across the realistic range and the toll is huge: moving from an ideal 50°F to a sweltering 80°F adds about 31 minutes to the median runner's day. In the famously hot 2012 race, the median finisher came in around 4:16, versus roughly 3:44 in cool, fast years.

Bar chart of median finish time by year, colored and labeled by temperature — Median finish time by year, shaded and labeled by race-day temperature. The hot years (2004, 2012) stand out in red.

The findings

Race-day temperature and the field's median finish time are strongly correlated: r = 0.86.
The cost is about 1 minute per °F on the median finish.
Going from a cool 50°F to a hot 80°F adds roughly 31 minutes.
The two hottest years (2012 at 81°F, 2004 at 74°F) were among the slowest; cool years (2009–2011) were the fastest.

Why it matters

If you're chasing a time goal — a Boston qualifier, a personal best — the forecast may matter as much as your training block. A warm race isn't a small handicap; for the average runner it can be the difference between hitting your goal and missing it by half an hour. Plan your goal pace around the weather, not just the calendar.

Data & method: Official Boston Marathon results 2001–2014 (public mirror: llimllib/bostonmarathon); race-day temperatures from the free Open-Meteo historical archive (mean of the 10:00–16:00 window near the course). Median taken over finishers between 2 and 8 hours. 2013 is excluded from the fit because the race was stopped (Boston Marathon bombing), truncating the finisher field. With 14 seasons this is a descriptive relationship, not a controlled experiment — field size and qualifying standards also shifted over the years — but the heat signal is large and consistent with sports-science studies. Reproduce with fetch_data.py + analyze.py in the repo.