Strategy7 minMay 5, 2026

Why 87% of Marathoners Run the Second Half Slower — And What to Do About It

Analysis of 300,000 Boston Marathon splits reveals exactly how much runners slow down, and why. The data points to one clear fix.

🔬
RunDataLab Research Team
Analysis backed by millions of race results

The marathon wall is not a myth. It shows up in the data with remarkable consistency.

We analysed split data from 300,000 Boston Marathon finishers across 14 years of race records, plus findings from published sports science research. The finding is stark: roughly 87% of runners complete the second half of the marathon slower than the first. And the slower your goal time, the worse the collapse tends to be.

What the Splits Actually Show

At Boston, athletes record split times at 10K, half marathon, 30K, and finish. This lets us calculate exactly how much each runner slowed from the first half to the second.

The result: positive splitting (running the 2nd half slower) is not just common — it's the overwhelming norm.

Finish Time Bracket% Who Positive SplitAverage Slowdown
Sub 3:0038%+2:30
3:00–3:3058%+5:00
3:30–4:0072%+9:30
4:00–4:3082%+15:00
4:30–5:0088%+22:00
5:00+91%+31:00

The pattern is clear: the slower the runner, the worse the blow-up. A 4:30 finisher on average ran the second half 22 minutes slower than the first — the equivalent of stopping for an extended walk break around mile 21.

🎯

Key Takeaway

The data shows that positive splitting gets dramatically worse as goal times increase. If you're targeting 4:00–4:30, your second half will be, on average, 15 minutes slower than your first unless you run a deliberately conservative first half.

Why This Happens: The Glycogen Cliff

The physiological mechanism behind the wall is well understood. At marathon pace, the body relies heavily on glycogen (stored carbohydrate) for fuel. Most runners carry roughly 90 minutes of glycogen at race effort. Once depleted, the body switches to fat burning — which produces energy far more slowly.

The result: a dramatic, involuntary slowdown. It's not a lack of mental toughness. It's chemistry.

The data shows this cliff appears most commonly between kilometres 32–37 (miles 20–23). Runners who hit the wall here see pace drop by 30–90 seconds per kilometre within just 2–3 kilometres.

⚠️
Warning

Boston's Heartbreak Hill (miles 20–21, right at the start of the glycogen depletion zone) makes this effect more severe than most marathons. The actual onset of slowdown in flatter races like Berlin or Chicago tends to occur slightly later, around km 35–38.

The 1% Rule: What the Fastest Runners Do Differently

Elite and sub-elite runners who avoid the wall share a common approach: they run the second half only marginally slower — or occasionally faster — than the first.

Analysis of Boston winners and top-10 finishers shows the optimal strategy: target a positive split of 0–3% between first and second half. In practical terms, for a 3:30 goal:

  • First half target: ~1:44:30
  • Second half target: ~1:45:30 to 1:46:30
  • Total: 3:30:00–3:31:00

Going out even 2 minutes too fast in the first half (1:42:30) dramatically increases the risk of a severe blow-up. The data from 300,000 races backs this up: there is almost no such thing as "banking time" in a marathon.

Pacing Bands by Goal Time

Here is what the data suggests as first-half targets to hit your finish goal while minimising blowup risk:

Goal FinishFirst Half TargetRecommended Split Differential
Sub 3:001:28:30–1:29:300–2% (0–2 min slower 2nd half)
Sub 3:301:44:00–1:45:001–3% (+1:30 to +3:00)
Sub 4:001:58:30–1:59:301–3% (+2:00 to +3:30)
Sub 4:302:13:00–2:14:002–4% (+3:00 to +5:00)
Sub 5:002:28:00–2:29:002–4% (+3:30 to +6:00)
🎯
Pro Tip

The counterintuitive finding: runners who go out 3–5% slower than target pace in the first half are significantly more likely to achieve their goal time than those who go out at exactly target pace. The safety margin matters.

The Fuelling Variable

One factor the split data can't fully separate: fuelling. Runners who successfully take on carbohydrates throughout the race show meaningfully better second-half splits than those who don't.

The research consensus points to consuming 60–90g of carbohydrate per hour from around the 45-minute mark. In practice, this means taking a gel or equivalent every 20–30 minutes from mid-race onwards.

Combined with a conservative early pace, this is the closest thing to a guaranteed wall-prevention strategy the data supports.


Data sources: llimllib/bostonmarathon GitHub dataset (300,000+ finishers, 2001–2014); SCORE Network 2023 Boston Marathon dataset; academic literature on marathon pacing and glycogen metabolism.