Module 2, lesson 2 · 40 minutes

Three ways a crowd kills: crush, collapse, stairway pile

The cause of death is nearly always the same, but force reaches the chest in three recognisable ways, each with its own trigger, geometry and defence.

You’ll be able to

distinguish a standing crush, a progressive collapse and a stairway fall-pile;
match each mechanism to a historical case;
explain the “hole in the crowd” domino mechanism, and why the people at the bottom of a pile die.

Three scenes. A terrace pen at a football ground, full, with more people arriving at the back. A narrow alley sloping downhill, packed from wall to wall, with people still trying to go both ways. A staircase at the end of a match, with thousands of people walking down it.

In all three, people died of the same thing you met in the last lesson: compressive asphyxia. But the way force reached their chests was different in each, and so is what would have prevented it. An expert reading an incident report wants to know first which of these mechanisms was at work, because that tells them where to look for the failure.

The three mechanisms. In a standing crush people are upright and the force comes horizontally through the crowd. In a progressive collapse a fall opens a hole that others are pushed into. On a stairway, gravity and the crowd behind keep feeding people onto a fall. Schematic, not to scale.

Crush: dying at the narrowest point

A standing crush (often just “crowd crush”) is compression of an upright, densely packed crowd against a barrier, wall or bottleneck, as force from people behind accumulates at the front. Helbing and Mukerji describe the classical form in one sentence: “In a classical crush, people are moving towards a physical bottleneck and die in front of its narrowest point”.

This is the mechanism of lesson 1. Arrivals keep coming from behind, the density at the front rises past the point where pushes can be absorbed, and the force adds up row by row on the people against the barrier. The victims are mostly upright. At Hillsborough, the Taylor report describes people “held upright by pressure all round”.

The trigger is almost always the same: more people arriving at a place than it can hold or pass. The narrowest point might be a pen front, a gate, a turnstile bank or the end of a corridor. The defence is therefore about flow: limiting inflow so that density never reaches the critical range, and keeping exits from that space open and working.

Collapse: the hole in the crowd

The second mechanism can happen where there is no barrier at all. Helbing and Mukerji contrast it with the classical crush: “In a ‘crowd quake’, there is typically no systematic flow directions, but people are pushed around by fluctuating forces”. Lesson 3 looks at those fluctuating forces in detail. Here, what matters is what happens when someone falls.

A person in a very dense crowd stumbles and goes down. That leaves a gap, a “hole” in the crowd. The people around the hole are still pushed from behind, but there is nobody in front of them any more to push back. They topple into the gap. Their fall opens a larger hole, and the next ring of people topple too. In Helbing and Mukerji’s words, those around the hole are “pushed from behind, but not anymore from the front… fall one after another like dominos, causing a pile”, and the people at the bottom “suffocate… under the weight of others”.

The domino mechanism. A fall removes the support in front of the neighbours, who are still pushed from behind; they fall into the gap, and the hole spreads. Schematic.

Progressive crowd collapse is Still’s descriptive term for a dense crowd falling in a spreading wave, so that people end up piled on top of one another. He uses it about the Hajj: “In ALL incidents at Jamarat the crushing is a progressive crowd collapse caused by the sheer numbers of pilgrims”. After Astroworld he explained that a surge leads to “a progressive crowd collapse… So the crowd actually falls on top of each other”. Note that it is a description, not a formally defined term: we found no formal definition from Still or anyone else that separates “crush” from “collapse”.

Itaewon, 2022

The alley beside the Hamilton Hotel in Itaewon, Seoul, is about 45 m long and 3.2 m wide at its narrowest, and slopes down towards the main road. On the night of 29 October 2022 it was packed with Halloween crowds. The police and forensic reconstruction describes a progressive fall: a first fall at about 22:15, four more falls behind it, and pushing from the top of the alley continuing for about 10 minutes. Deaths were from crush asphyxia or brain swelling.

How dense was it? A police and National Forensic Service simulation gives 10.74 p/m² at 22:25; an AP report gives about 8 p/m² at 22:15. The two figures are for different moments and come from different methods, so treat them as a range rather than a contradiction. Either is far above the 6 p/m² where, as you saw in lesson 1, a crowd stops absorbing force.

Someone falls in the middle of a dense, stationary crowd on a flat square. Explain why people who did not trip fall too, and why the people at the bottom are the ones most likely to die.

The fall leaves a hole. The neighbours are still being pushed from behind but are no longer supported from the front, so they topple into it, which opens a larger hole for the next ring. Nobody needs to trip: the forces that were balanced while everyone stood are unbalanced as soon as one person is missing. The people at the bottom carry the weight of those on top and the push of the crowd behind, cannot expand their chests, and die of compressive asphyxia.

Stairway pile: gravity joins in

A stairway fall-pile is a fall on a stair that the crowd behind cannot stop for, so that people fall onto those already down and a pile builds. On a stair the domino mechanism gets help. People coming down cannot see a fall clearly, cannot step over it, and are pushed on by the crowd behind and pulled on by gravity. Every step down makes the pile higher.

Ibrox, Glasgow, 2 January 1971. 66 people died on Stairway 13 as the crowd left a Rangers v Celtic match. One or two people fell, and continued pressure from above built a pile; most deaths were from compressive asphyxia. A popular story said fans had turned back up the stairs after a late goal. The Fatal Accident Inquiry rejected it: everyone was moving the same way. There had been earlier deaths on the same stairway in 1961, and incidents in 1967 and 1969.

Fruin estimated what the bottom of such a pile feels like. Assuming a pile 3 m high, “people on the bottom would experience chest pressures of 3600-4000 N (800-900 lbs.)”. That is his own calculation, not a measurement, but it puts the bottom of a pile well above every fatal figure in lesson 1’s table.

Bethnal Green tube shelter, London, 3 March 1943. The trigger was an anti-aircraft rocket salvo. A woman and a child fell near the bottom of a dark stair, and a pile of about 300 people formed; 173 died (some sources say 178). Afterwards the steps, the lighting and a central handrail were changed: each one aimed at a different part of the mechanism.

Luzhniki, Moscow, 20 October 1982. Another stairway pile, with 66 deaths by the official count published in 1989.

On a level floor, a fall in a slow crowd can sometimes be absorbed: people stop and help. On a descending stair, the people above cannot stop quickly, cannot see, and are pushed on from behind, so a single fall near the bottom can grow into a pile of hundreds. Handrails, lighting, even steps and control of how many people enter a stair at once all attack this mechanism.

Mechanisms mix

Real disasters do not come labelled. At Hillsborough some victims were held upright by pressure all round, a standing crush; then a barrier broke and “Many fell and the involuntary rush of those behind pressed them down”, a collapse. At the Love Parade in Duisburg in 2010, the crowd was crushed on a ramp that was the only way in and out, and video shows people who “stumbled and piled up due to a ‘domino effect’”.

The classification is still worth having. It tells you which trigger to look for and which defence failed. Module 8 uses this table for every case study.

Three mechanismsThe cause of death is the same in all three: compressive asphyxia. What differs is how the force arrives.
Mechanism	Typical trigger	Geometry	Victims	Example	First line of defence
Standing crush	Inflow greater than the space or exit can take	A barrier, wall, gate or bottleneck in front of a dense crowd	Mostly upright, at the front	Hillsborough 1989 (pens 3 and 4), with collapse later	Control inflow and capacity; keep exits from the space open
Progressive collapse	A fall in a crowd above the density where force can be absorbed	Open ground or a passage; often a slope; no single barrier needed	Fallen, in a pile	Itaewon 2022; Love Parade 2010	Keep density below the critical range; stop inflow and counterflow
Stairway fall-pile	A fall on a stair with people still coming down	A stair, often dark, steep or crowded from above	Fallen, at the bottom of the pile	Ibrox 1971; Bethnal Green 1943	Handrails, lighting, even steps; meter the flow onto the stair

Classify this: a crowd leaving a stadium is funnelled into a steep, poorly lit stair. Near the bottom, someone trips. What mechanism do you expect, and name two changes that would each attack a different part of it.

A stairway fall-pile. Good lighting and a central handrail make a fall less likely and give people something to hold; metering the flow at the top of the stair (fewer people on it at once) limits how many can be pushed onto a fall. Even step dimensions also reduce trips. Bethnal Green’s post-disaster changes, to the steps, the lighting and a central handrail, are exactly this list.

“A crush needs people running.” A standing crush happens in a crowd that is barely moving. The force comes from pressure, not speed.
“Stairway disasters happen because people rush or turn back.” At Ibrox the inquiry rejected the story that fans turned back for a late goal: everyone was moving the same way.
“Each disaster has one mechanism.” Hillsborough had upright and fallen victims; the Love Parade had a crush and a domino collapse.
“Progressive crowd collapse is a defined technical term.” It is Still’s descriptive phrase. No formal definition separating crush from collapse has been found.