Data story · 31,912 marathon finishers

The Negative-Split Myth

Every coach says to run the second half faster than the first. So how many marathoners actually pull it off? I checked an entire Boston field.

By Jeremy Lee

2.5%

of finishers ran a true negative split

+14%

the median runner's second half was slower

4×

over-4h runners slow ~4× more than sub-3h runners

The advice everyone gives

"Negative split" is the most repeated piece of marathon wisdom there is: hold back early, finish faster, and you'll run your best time. It's great advice. The question I wanted a real number for is simpler — out of thousands of ordinary runners, how many actually do it?

What I did

I took the official results of the 2014 Boston Marathon — 31,912 finishers with timing-mat splits — and split each runner's race in two: the time to the half-marathon mark, and the time from there to the finish. If the second half was faster, that's a negative split.

Almost nobody speeds up

The verdict is blunt. Only 2.5% of finishers ran a negative split. Another ~4% paced within 2% of even. The other ~94% slowed down — the typical runner's second half was 14% slower than the first.

Bar chart: 2.5% negative split, ~4% even, ~94% positive split — The overwhelming majority of the field ran a positive split (slowed down).

Histogram of how much slower each runner's second half was — How much slower the second half was for each runner. The whole distribution sits to the right of "even pace."

The better you are, the more evenly you run

The slowdown isn't the same for everyone. Sub-3-hour runners gave up only about 5% in the back half; runners over 4 hours lost more than 21%. Even pacing isn't just good advice — it's a signature of fitness and experience.

Line chart: median slowdown rises steadily from fast to slow finishers — Median second-half slowdown by finish time. Faster finishers pace far more evenly.

A small gender gap

Women paced slightly more evenly than men — a median slowdown of about 13.4% versus 14.8% — echoing a pattern other researchers have found across many races.

Bar chart comparing median slowdown for men and women — Median slowdown by gender, with the share running a negative split.

The findings

Only 2.5% of 31,912 finishers ran a true negative split; ~94% slowed down.
The median runner's second half was 14% slower than the first.
Pacing tracks ability: sub-3h ≈ +5% slowdown vs over-4h ≈ +22%.
Women paced a touch more evenly than men (≈13.4% vs ≈14.8% median slowdown).

So is the advice wrong?

Not at all — it's just hard. The data says holding back early is exactly what separates the strong finishers from the strugglers, and almost no one errs on the side of going out too slow. If anything, the takeaway is humbling: next time, start slower than feels right. You're very unlikely to overdo it.

Data & method: Official 2014 Boston Marathon results (public mirror: llimllib/bostonmarathon), 31,912 finishers with valid half splits. First half = half-marathon split; second half = finish − half split. Reproduce with fetch_data.py + analyze.py in the repo. One race-day caveat: weather and course (Boston's downhill start, late uphills) shape the exact numbers, but the lopsided pattern holds everywhere it's been studied.