Even pacing

Why does an even pace beat a fast start over a long race?

Two runners cover the same distance in the same total time. One runs it at a steady pace throughout. The other goes out fast, banks a lead, and hangs on. Their averages are identical, so on paper nothing distinguishes them.

Yet coaches are near-unanimous that the second runner has run the slower race — that had they paced evenly, they would have beaten their own time. Why should the distribution of effort matter at all, if the total is the same?

Start by noticing that "the same average" quietly assumes the cost of running is proportional to the speed. If it were, the two strategies would indeed be equivalent, and the whole debate would evaporate. So the question becomes: is the physiological cost of intensity linear?

Consider what changes when you run harder. Below a certain intensity, the rate at which your muscles produce metabolites is matched by the rate at which they clear them, so a steady internal state is possible and the effort can continue for a long time. Push past that ceiling and production outruns clearance. There is no new steady state; there is only accumulation, and a finite amount of accumulation you can tolerate before you must slow down.

That single fact changes everything, because it means intensity above the ceiling draws on a store rather than a rate. A useful two-parameter description used in exercise physiology captures it: performance is described by a sustainable ceiling — critical power, or critical speed — plus a fixed capacity for work above it. Spend that capacity and you are forced back to the ceiling or below.

Now do the arithmetic that this implies, because it is where the intuition fails. If you have a fixed capacity to spend and you spend it at a certain rate, the time you can hold the effort is inversely proportional to how far above the ceiling you are. Go twice as far over, and you last half as long. Not ten percent worse — half. The cost of intensity is not linear but curved, and steeply.

Do you see what that does to the fast start? The seconds gained in the first mile are bought at a much higher metabolic price per second than the seconds lost in the last mile are refunded. You are buying at the expensive end of the curve and selling at the cheap end. Even before physiology adds anything else, a convex cost curve punishes variability in the effort for any fixed average.

And physiology does add more. Running above the sustainable ceiling recruits less efficient muscle fibres and drives the oxygen cost of the same speed steadily upward over time — the so-called slow component — so the fast start makes the remainder of the race literally more expensive per metre than it would otherwise have been. The damage is not merely stored; it compounds.

Is there a check that does not depend on the model? Look at how records are actually run. In distance events from around 1500 m upward, the fastest performances in history are run with close to even splits, frequently with a slightly faster second half. Nobody sets a world record by going out hard and hanging on.

Several qualifications, because "even pace is optimal" is a good rule and not a law. The two-parameter model is a useful approximation rather than a full description: the capacity above the ceiling can be partially restored by dropping below it, and the parameters themselves drift over a long event.

The rule also weakens as races get short. Over the 800 m the deficit never fully comes due before the finish, and the fastest performances are run with a first lap markedly quicker than the second — the current world record by more than two seconds. Sprints are all overshoot. So the correct statement is that even pacing matters more the longer the event, because there is more race left in which to pay.

Tactics can also override physiology. A championship final is a race to place, not to a time, and reacting to a break may be worth more than the metabolic efficiency it costs. Terrain matters too: an even effort over hills is not an even pace, and it is effort, not speed, that the argument is really about.

Finally, a caveat about explanation. Pacing is often said to be regulated by an anticipatory process — teleoanticipation, or the central governor model — in which the brain sets output with reference to the distance remaining, so the athlete slows before catastrophic failure rather than because of it. Something like anticipatory regulation is clearly happening, since athletes routinely produce an end-spurt from a state they had called maximal. But the central governor framing specifically remains contested, and the argument here does not depend on it: the convex cost of intensity is enough on its own.

• How much of the capacity above the ceiling can be recovered by easing off mid-race, and how fast.
• Why an end-spurt is possible at all, if the athlete was already at their limit.

Every term the collection defines is gathered in the glossary.