Every runner faces the same question on the start line: do I go out conservatively and speed up (negative split), hold one steady pace from gun to tape (even split), or bank time early in case I fall apart later (positive split)? The wrong answer can cost minutes you can’t get back.

This guide explains what each strategy actually means, what the data says about which works, and how to execute the right one for your race, your fitness, and your course.

Definitions

Even split: the second half of the race is run in the same time as the first half (within ~1%). Often called “perfect pacing.”
Negative split: the second half is run faster than the first. A “2% negative split” in a 4:00 marathon means a first half of 2:01:12 and a second half of 1:58:48.
Positive split: the second half is slower than the first. The default outcome for most recreational marathoners.
Even effort vs even pace: on hilly courses, “even effort” means the same perceived exertion the whole way, which produces uneven pace. On flat courses, even effort and even pace converge.

What the data says

Across published analyses of major-marathon finisher data, the pattern is consistent: the fastest runners run close to even or slightly negative; the slower the runner, the larger the positive split tends to be. World-record marathon performances over the last two decades have nearly all been run within 1-2% of even pace, and many of them with a slightly faster second half. Among recreational marathoners, the median positive split is in the range of 5-10%.

The lesson isn’t that negative splits are magic. The lesson is that most runners start too fast, and the cost of doing so compounds over the second half of the race. The “fastest split distribution” and “lowest blowup rate” are essentially the same strategy: don’t go out too hard.

The physiology of why positive splits hurt

Three things happen when you start a marathon 10-15 seconds per kilometer faster than your sustainable pace:

Glycogen depletion accelerates. The faster you run early, the more carbohydrate you burn (vs. fat) per minute. You hit “the wall” — bonking — earlier and harder.
Lactate accumulates. Even small early surges can push you above your lactate threshold. Once lactate accumulates, clearance during sustained running is very limited.
Core temperature rises faster. Heat production scales with running speed. A 5% start-too-fast in warm conditions can mean a 15-20% slowdown in the second half.

By contrast, starting 5-10 seconds per kilometer slower than goal pace costs roughly 30-60 seconds in the first 10 km but typically returns 2-4 minutes in the back half. The trade is enormously favorable.

When to even split

Even splitting is the default recommendation for most runners on most courses. It’s simple to execute, it minimizes pacing errors, and the data backs it up. Choose even splits when:

The course is flat (Berlin, Chicago, Valencia, most local marathons).
Conditions are stable (no big weather change forecast mid-race).
You’ve done your homework and you trust your goal pace.
You’re an experienced runner with multiple marathons of pacing data.

When to negative split

A planned negative split — running the second half intentionally 1-3% faster than the first — is appropriate when:

The course’s hard half is first. NYC’s bridges and Central Park hills make a sub-3:30 effort almost impossible to even-split; a 1-2% positive split run by feel is essentially the same as an even-effort race. See our NYC Marathon pace guide.
You’re racing in heat that will rise during the race. Counterintuitively, this argues for going out slower than goal pace, not faster.
It’s your first race at this distance. A planned conservative start is the cheapest insurance against blowing up.
You’re targeting a specific BQ or sub-X time and you want to maximize the probability of finishing under that mark, not the expected finish time.

When positive splits are unavoidable

Positive splits are not always a tactical mistake. Sometimes they reflect the course honestly:

Net-uphill second half: some courses are simply harder in the back half, and an even-pace strategy is impossible.
Boston-style downhill openings: Boston’s first half is net downhill. Running an “even effort” produces a positive split in time, and that’s actually correct. See our Boston Marathon pace guide for the specifics.
Heat that rises sharply mid-race.

The question isn’t “did I positive split?” — it’s “did I positive split because of the course/conditions, or because I went out too fast?” The first is a choice; the second is an error.

Even pace vs even effort on hilly courses

On flat ground, the two are the same. On hills, they diverge. Running an even pace up a hill requires far more effort than running the same pace on flat. Most experienced racers run even effort, accepting the slowdown on uphills and the speed-up on downhills. The grade-adjusted pace concept formalizes this: a 5:30/km pace up a 5% grade equates to roughly 4:50/km on flat ground. Race-specific calculators (NYC, Boston, etc.) bake this in.

How to execute a planned conservative start

The hardest part of a smart pacing strategy is the first 5 km. You’ll feel fresh, the field will surge around you, and your watch will say you’re going too slow. Stick to the plan.

First 2 km: 10-15 sec/km slower than goal pace. Watch is your honesty cop, not your competition.
km 2-10: drift toward goal pace. By 10 km, you should be at goal pace exactly.
km 10 to halfway: goal pace, locked in.
halfway to 30 km: goal pace, with a small mental note: do I feel strong? Lock in or back off accordingly.
km 30+: if you’ve executed the first 30 km correctly, you have permission to push.

Common pacing mistakes

“Banking time” early. The most expensive mistake in marathon racing. The deficit you create by banking 60 seconds in the first 10 km typically costs 3-5 minutes in the last 10 km.
Ignoring the course profile. Pacing a hilly course like a flat one is a recipe for blowing up on the first big climb.
Pacing off your watch’s auto-lap pace alone. GPS pace is noisy in the first kilometer (skyscrapers, tunnels) and during turns. Trust your perceived effort and your overall split, not the instantaneous reading.
Surging at every aid station. Drinking from a cup costs 5-10 seconds. Trying to “make it back up” after the station shreds your pacing.

What this means for your training

Practice your race-day pacing strategy in your long runs. A great workout six weeks out: 28 km with the first 8 km deliberately slow, the next 16 km at goal marathon pace, and the last 4 km moderate. You’ll teach yourself what “going out easy” actually feels like — and the answer is “this feels too easy.” That’s how it’s supposed to feel.

Map your splits with the pace calculator, and pull race-specific grade-adjusted splits from any of our race calculators.

FAQ

Should I always run a negative split? No. On flat courses with stable conditions, an even split is the gold standard. Negative splits are most valuable when there’s an asymmetry — course, weather, or risk tolerance — that justifies them.

How much slower should I start? For most marathons, 5-10 seconds per kilometer slower than goal pace for the first 2-3 km is enough. More than that and you give back too much; less than that and you risk going out at goal effort, not goal pace.

Why do elites run faster than even sometimes? Elite marathon races are often paced for a record attempt, with the second half intentionally faster than the first. Their version of “negative split” is a few seconds per kilometer; for recreational runners, the same concept becomes a margin against blowing up.

Does pacing matter for shorter races? Yes, but less. In a 5K, a 5-second positive split is normal and recoverable. In a marathon, a 5-second-per-km positive split means a 4-minute slowdown over 26.2 miles.

Can I negative-split a Boston Marathon? Practically, no. Boston’s downhill first half makes a true negative split nearly impossible without sandbagging the opening. The right model for Boston is “controlled positive” — let gravity help, don’t fight it, but don’t bank time either.

What about pacing for a half marathon? The same principles apply but the margin is smaller. A 5-second-per-km positive split in a half costs roughly 1:45; in a marathon it costs 3:30. Half-marathon-specific advice: target the same goal pace from km 2 onward, with the first 1-2 km 5-7 sec/km easier.

How do I know my goal pace? Use our race predictor. Input a recent race result and it projects your equivalent times at every other distance. The marathon prediction in particular tends to be optimistic without specific marathon training; for goal-setting, take the predicted time and add 5-10% if you haven’t done a long-run block.

One more thing: the watch trap

The most underrated cause of bad pacing is a panic glance at the watch in the first kilometer. GPS pace is noisy in the opening minutes (skyscrapers, tunnels, dense crowds). Your watch may show 5:25/km when you’re actually running 5:38; the natural reaction is to slow down, then speed back up. Both moves are wrong. Trust your perceived effort for the first 2 km and let the watch settle.

The same trap appears in the late race in reverse. At km 35, watch pace lags behind real effort because legs are heavier and stride is shorter; runners panic-surge to “make up” pace and blow up at km 40. The right reaction is to keep effort steady; the watch is telling you you’re tired, which you already knew.

Negative Split vs Even Split: Which Race Pacing Strategy Wins?.