Race pacing
Evidence: moderate
Even or slightly negative pacing wins most distances longer than two minutes, and a fast start reliably costs more than it gains. The case rests largely on observational race data and physiological modelling rather than randomised trials, so the headline is sound but the precise optimum is not pinned down.
How you spread effort across a race changes the result. The same fitness produces a different finish time depending on whether you run even, start fast and fade, or start steady and accelerate. A foundational review classified the main profiles as all-out, positive, even, negative, parabolic and variable, and matched them to event duration (Abbiss & Laursen 2008). The practical message from that work is simple. Very short events reward going all-out, and events longer than about two minutes reward distributing pace more evenly (Abbiss & Laursen 2008).
Why a fast start costs more than it gains
The energetic argument is the core of it. Holding a steady speed avoids wasting energy on repeated accelerations, so uneven pacing tends to raise the metabolic cost of covering the distance (Abbiss & Laursen 2008). This curvature-of-the-cost-curve argument is strongest in cycling, where air resistance makes power rise steeply with speed; in running the cost of speed is much closer to linear, so for distance runners the larger penalty of a fast start is physiological rather than mechanical. A surge early in a race is not refunded later. It is spent at a premium, because the extra speed draws disproportionately on anaerobic energy and accelerates the depletion of glycogen.
Glycogen is the binding constraint in long races. Running faster than goal pace early burns a larger fraction of carbohydrate, and once stores run low late in a marathon the body must lean on fat oxidation, which supplies energy at a lower maximum rate than carbohydrate (Grivas 2025). Conservative early pacing spares glycogen, blunts early lactate accumulation, attenuates cardiovascular drift, and limits heat build-up, all of which delay the fade (Grivas 2025). The cost of a fast start is therefore not linear. Small overshoots produce modest slowing, larger overshoots produce a collapse out of proportion to the time banked early. See in-race carbohydrate and durability for the fuelling and fatigue-resistance side of the same problem.
What the race data show
Elite marathoners run remarkably flat. Analysis of championship marathons found the best runners varied their speed by only about 3% across the whole race, and runners who held an even effort fared better than those who drifted off the pace (Hanley 2016). Recreational runners do the opposite. The large majority record positive splits, slowing in the second half, and only a small minority manage a true negative split (Hanley 2016). The gap between the two groups is largely a pacing-discipline gap, not only a fitness gap.
A caution about the evidence. Most of this comes from observing what fast people do, not from randomly assigning runners to pacing strategies. The even-pace pattern of elites could partly reflect that the most capable runners both pace well and run fast for other reasons. The physiological case for even pacing is strong and the race data are consistent with it, but the design is observational, which is why this page grades moderate rather than strong.
Pacing by distance
Practical guidance
- 800 m and shorter. The first lap is almost always faster than the second. This is partly an unavoidable artefact of accelerating from a standing start and of positioning, and partly because top-end speed cannot be held once lactate climbs (Abbiss & Laursen 2008); whether a positive split is genuinely optimal or merely typical is less settled. In practice, commit early and hold on.
- 1500 m to 10 km. Aim for even effort. Avoid the adrenaline-fuelled fast first lap, settle into goal rhythm quickly, and save any change of pace for the finish.
- Half and full marathon. Start at or just below goal pace and aim for even or slightly negative splits. The classic error is banking time early; it is almost always withdrawn with interest after 30 km (Grivas 2025).
- Ultras and hilly courses. Hold constant effort, not constant pace. Over hills, keeping exertion steady rather than speed steady is the more economical strategy, so ease on the climbs and let the descents come back to you (Abbiss & Laursen 2008).
Conditions
Heat, wind and terrain all justify pacing slower than a flat-calm goal. Hills are best run as constant effort. Headwinds are worth tucking behind others, and championship data show that running in a pack with similar-ability runners helps both sexes, while running alone tends to cost time late (Hanley 2016). The discipline to start conservatively is the single most transferable pacing skill, and it is best rehearsed in training so race-day adrenaline does not override it.