Anatomy of a mass-participation race

Introduction

In April 2026, 59,830 people finished the London Marathon, a Guinness World Record for the largest marathon, beating the 59,226 who had finished New York the previous November. Seven years earlier, London had 42,549 finishers. That is about 40% more runners, and none of the roads got 40% wider.

A marathon is an unusual crowd. You know almost exactly who is coming, roughly when each of them will start and, from the finish time they predicted when they entered, roughly how fast each will go. That makes it the most plannable crowd there is. This lesson takes a race apart so that, by lesson 3, you can plan one.

The biggest races keep getting bigger

The six long-established World Marathon Majors became seven when Sydney joined; its first race as a Major was on 31 August 2025. Table 1 shows how large the biggest of them now are. Several numbers come with a caveat, which is normal for this kind of figure: race reports, record bodies and magazines count starters and finishers differently and sometimes simply disagree.

Fifty or sixty thousand people is the population of a market town, all moving the same way along streets that were built for traffic.

The parts of a race

Figure 1 lays out the parts of a big race that matter for crowding. Start at the left.

• Waves and corrals. A field this size cannot start at once. It is split into waves: groups released across the start line one after another, each starting a set time after the previous one. Before the start, runners wait in corrals: fenced pens where the runners of one wave, or one pace group within a wave, gather before their start. London bases its waves on runners’ predicted finish times. A runner may move to a later wave but not to an earlier one.
• The start line. However wide the road, only so many runners a minute can cross the line. This start-line throughput is the number of runners per minute who cross the start line while a wave is being released. It is the single most important number in this module.
• The width profile. A course is a sequence of roads, bridges and junctions, each with its own usable width. Planners think of it as a width profile: width against distance.
• Pinch points. A pinch point is a short stretch of a route that is much narrower than the road either side of it, such as a bridge, a gate or a narrow street. Boston’s course begins in Hopkinton, where Main Street “squeezes to just 39 feet wide”, which is about 11.9 m.
• Aid stations. Runners slow down, swerve or stop to drink, so the same flow of runners occupies more road.
• The finish funnel. The road narrows into chutes where runners stop running, collect medals and bags, and need medical attention. Everyone arrives on foot, but they leave much more slowly than they came.

The hidden variable: the spread of paces

Runners do not move as a block. Each has their own pace: time per kilometre, in min/km; a runner’s predicted finish time divided by 42.195 km. If a runner crosses the start line at time s and holds pace p, they reach distance x at time

Two runners who cross the line together at 5 and 6 min/km are 10 minutes apart at 10 km and 42 minutes apart at the finish. The gap grows in proportion to the distance run. So a field with a wide spread of paces stretches out along the course and gets thinner as it goes. Figure 2 shows it.

Worked example: runners per minute at a point

10,000 runners cross the start line evenly over 10 minutes, 1,000 a minute. Their paces are spread evenly between 5 and 7 min/km.

1. At 10 km the first arrives at 0 + 5 × 10 = 50 minutes and the last at 10 + 7 × 10 = 80 minutes. 10,000 runners over 30 minutes is about 330 a minute.
2. At 40 km: first at 0 + 5 × 40 = 200 minutes, last at 10 + 7 × 40 = 290. 10,000 over 90 minutes is about 110 a minute.

The start line saw 1,000 a minute. The 10 km point sees a third of that and the 40 km point a ninth. The field is at its most compact just after the start. That is why the start, and the first pinch point after it, is where crowd planning for a race begins.

The exception is a field that is started in the wrong order. If slower runners are sent off in front of faster ones, the fast runners close the gap and the field bunches mid-course. You will see exactly that in the wave planner.

From runners per minute to space per runner

A flow of runners becomes a density once you know how fast they are going and how wide the road is. This is the same flow equation as Module 1, lesson 2, with runners instead of walkers:

Worked example. At 10 km in the example above, about 330 runners a minute pass at around 6 min/km, which is 1,000 ÷ 6 ≈ 167 m a minute. Each metre of road holds 330 ÷ 167 ≈ 2 runners. On a road 10 m wide, that is 0.2 runners/m², or 5 m² each.

Now compare the start. 1,000 runners a minute at the same speed means 6 runners on every metre of road. On a 12 m road that is 0.5 runners/m², 2 m² each, and on a 5 m pinch it is 1.2 runners/m².

What counts as enough room for runners? The design target reported for New York is “three square meters for each runner”, about 0.33 runners/m². That is far more space than walkers need, which makes sense: a runner needs clear road in front of every stride that a walker does not.