Crowd Risk Analysis and Crowd Safety
April 8, 2015

We have all experienced crowds, some good, some bad and some very frightening. If you have ever been swept off your feet in a crowd crush you will know how powerless you can feel in that situation. In this article, Professor Dr. G. Keith Still outlines the principles and applications of Crowd Risk Analysis and crowd modelling for places of public assembly. 

Crowd Risk Analysis

Over the last twenty-five years, we have been developing a practical approach to assessing the risks related to“Crowd Dynamics[1]”. The objective of this work is to define a hands-on methodology for risk assessment and to assist users in the recognition of risks related to overcrowding.

 What are the risks of crowd crushing?

John Fruin wrote a paper in 1993 calledCauses and Prevention of Crowd Disasters[2]”. In this paper, published in 1993[3], he describes the underlying causes of crowd related accidents and incidents. Specifically, Fruinhighlights that the forces that can be exerted by a 3-5 people

  • In an Australian Building Technology Centre study, three persons in a combined leaning and pushing posture developed a force of 1,370 N (306 lbs.). This study showed that five persons were capable of developing a force of 3,430 N (766 lbs.).

In the same book, a paper by Hopkins called “Crowd Pressure Monitoring[4]” outlines dangerous crowd pressures:

  • Firstly, death was estimated to have occurred 15 seconds after a load of 1,440 lbf (6,227N) was applied. ….it was estimated that death occurred after 4 – 6 minutes of applying a load of 2,50lbf (1,112N).

We can see that it does not take a vast crowd to exert dangerous pressures.

Some of the key questions we are often asked during lectures and workshops are:

  • How many people can this space hold? What are the risks of crushing? How do we measure crowd risks during an event?

The following visuals will help define the safety criteria for the above questions.

Identifying risks of dangerous pressure

The images below show the crowd density from 1 – 5 people per square metre, shown against two tennis courts (625 square metres 5m grids):

Untitled-1 Untitled-2 Untitled-3 Untitled-4 Untitled-5

When we look down on the crowd, we get a clearer perspective of crowd density. Even at five people per square metre, we do not appear to have physical contact between people in the crowd. There can be no crowd pressure without this physical contact.


If you are planning an event, checking a plan, or involved in crowd management, you need to understand the risks of overcrowding; how and where overcrowding may occur and how to prevent dangerous overcrowding from developing. Maintaining low crowd density is critical to the principles and applications of crowd safety. This is a fundamental principle of crowd risk analysis. However, we need to clarify the difference between hazards and risks.

 Crowd Risk Analysis

When preforming a risk analysis (specifically the risk of crowd crushing) you need to understand the concepts of “hazards and risks”. A hazard is a potential source of danger; a risk is a situation involving exposure to danger. Therefore, if the crowd density increases above five people per square metre (hazard) there is a significantly increased risk to life and limb (from crushing). We observe a number of phenomena occurring above five people per square meter, such as crowd surging, crowd collapse (the domino effect) all of which can have lethal consequences if left unchecked. It is essential, when planning for mass gatherings, that there is a competent person reviewing the risks. Therein lies the key issue – how do we define a competent person? Would simulations help? How can we ensure crowd safety in places of public assembly?

We may never fully understand the mechanisms of fatal incidents, to do so would require extensive study and destructive testing of the crowd but we can define an approach to risk analysis and risk management, specifically the risks of overcrowding using a relatively simple set of analysis tools. To do this we need to understand the “redline”, the limits to crowd density in built and complex spaces.

The RedlineUntitled-7

Most vehicles are fitted with a tachometer or rev counter, which measures engine revolutions per second. The manufacturer will highlight (normally in red) the maximum speed at which the engine is designed to operate without causing damage. If you rev the engine above the Redline, the probability of failure increases significantly.

 Crowd density and crowd risk

The equivalent of the rev counter, and the “redline”, for the crowd dynamic is crowd density, which is measured as the number of people per square metre. If the crowd density increases above a safe limit, the risk of crowd crushing and subsequent injuryincreases significantly. There are, of course, different limits for maximum crowd density and for both the standing and moving crowds. To understand the principles, we have a series of simple experiments, we’ve run these in classrooms around the world and it can help a team understand the concepts of crowd risks, and specifically the risk of crowd surging. To illustrate this we run a simple experiment in the classroom, which demonstrates the risks to a standing and moving crowd using a loop of string with an internal area of 1 square metre.

Standing/static crowd risks

We mark out 1 square metre on the ground and invite people to step into the square, one at a time, until it becomes uncomfortable (around 5 people per square metre). Then we add a few more people to the square. When the group in the crowd are packed to the extent that they are all in physical contact (around 6-7 people per square metre) any movement has a ripple (or domino) effect. We demonstrate this with a small push on one side of group, which results in the person at the far side of the group feeling a much larger effect. As each person shifts following the small push, his or her momentum adds to the effect. This is called a “shockwave” and in crowds, this can be extremely dangerous. However, shockwaves only occur when the density exceeds the safe limit. We have run this experiment with adults and which students, in some cases we can get 9-10 students in the square, in other cases only 5-6 adults. You need to understand the demographics of your crowd when evaluating the density limits. The principles should be self-apparent, but we recommend further reading on this specific topic.

 Dynamic/moving crowd risks

There are a series of video clips on the website[1] that show the experiment for a moving crowd. Above three people per square metre, we note a significant drop in crowd speeds. From this we can define two redlines for crowd density risks.

Static (standing crowds) – more than five people per square metre

Dynamic (moving crowds) – more than three people per square metre

It is essential to regulate crowd flow and space filling so that the crowd is not maintained below the “redline” and not exposed to risk of overcrowding. This is the principle of crowd risk management, realising the risk, and then managing the environment to prevent the crowd being exposed to those risks.

 What about the front of stage area?

In areas, such as the front of stage at a concert, where the crowds will choose to pack to a much higher density, we need to monitor the crowd and have a contingency for assisting anyone in distress. This is the function of the front of stage pit crew.


The crowd is monitored and the “pit crew” would be trained on how to recognize distress in the crowd and act accordingly. These are mainly static crowds, observing and enjoying the performance. When evaluating a moving crowd we need to expand on the principles of risk analysis and site evaluation.

Risk analysis and site evaluation

The objective is to prevent a dangerous situation developing and for that we recommend a simple approach to understanding the “redline” risks. We call this a RAMP analysis (Routes, Areas, Movement and Profile). We will describe this process step-by-step.

RAMP analysis – routes

Using a wide area site plan, determine which routes the crowd will take to get to the event. Typically, you can achieve this by reviewing the local transportation system, information that you can obtain from a variety of mapping tools freely available on the Internet.

People will travel by train, bus, car or walk from local population centres. Determine where the transport terminals are located and evaluate the percentage of people using those routes from the transport system to the event. Prepare a routing diagram (illustrated below). Always walk the route, it helps you understand how people will move towards your event and can highlight areas where signage, information or stewards would assist the process of getting to your site.


This diagram shows the train station to event space (a football victory parade through Manchester, UK). We can see that one side of the route is likely to fill first (and that there may be problems of filling the other side of the route. We can also see there is s section of the route with no pubic transport access. The routing diagram has highlighted a few important issues of where we may expect higher crowd densities. This indicates capacity planning, routing and crowd management issues.

For example, there is a station very close the stadium (bottom left) we may need to consider closing this as the ingress (approaching crowds) will be limited by the train capacity, but the departing crowds will converge on the station and that could lead to dangerous overcrowding. This is redlining; we are not performing a detailed analysis, just reviewing the potential for overcrowding (how and when overcrowding may occur). The station near the stadium is too close, it may work for a stadium event, but there could be problems for a victory parade where people are gathered outside the stadium and along the route.

Further analysis is required, such as

  • How many people may use this station?
  • What may we need to manage the egress process?
  • How does egress function for a stadium event?
  • Will the victory parade crowd be larger than the stadium capacity?

That is the objective of redlining, to highlight the issues of this specific site, used for this specific mass gathering? One key element of this is to evaluate the available area.


There are a variety of freemeasuring tools available on the Internet.They allow the user to measure the area for the site from a wide range of maps,plans, diagrams or satellite images (such as Google Earth). Below we have illustrated a city centre that was planned to host a parade, with market stalls and have a moving crowd (in the order of 60,000 – 90,000) throughout the day. We can evaluate the available space (grey = available space, and we need to remove the areas in green = stalls and the yellow = parade route). This left only 7,400 square metres. A rough estimate is that 2-3 people per square metre (freely moving around the site) would give us a capacity or approximately 14,800 to 22,000 people. We were expecting 60,000 – 90,000 people, a clear indication of a redline, the space is not big enough, we have a potential problem with this proposal.


Static and dynamic spaces

To assess crowd risks, we recommend a site be divided into static and dynamic spaces, those areas where the crowd is generally standing (static spaces) and observing the event and those areas where the crowd is generally going to be moving (dynamic spaces).

There are event guidelines, which start with an evaluation of the total available area and then multiply this by 2 (for 2 people per square metre) this is a good starting point for the maximum capacity of the site. However, the event space will have predictable areas of high (static), low (dynamic) and underused (empty) spaces.

It is important to understand how the space is going to be used and if any areas of the site are going to experience localised high density (and hence increased risk). You simply mark out these different areas on the site plan and measure them using any of the available area measurement tools. We can now evaluate the site risks and mark the plan accordingly, no more than 5 people per square metre on static spaces (this value will depend on the demographics of the crowd, family groups, buggies, back-packs will need more space, less density) and no more than 3 people per square metre on the dynamic spaces. We recommend that site maps, plans and diagrams clearly indicate these areas for planning and licensing or site approval purposes. You will need to evaluate if backpacks, buggies, family groups, picnic hampers, etc. will have an impact on space utilisation and hence the crowd density limits.


Crowd safety is not only a function of how much space is being occupied, it also relates to how quickly those spaces may fill.Using the routing diagram, we can trace along the route and check any pinch points (or bottlenecks). By applying a “first pass” or rough cut capacity (RCCP) analysis we can assess the maximum flow rate though the system, which will be constrained by the flow at the narrowest point in the system.

The RAMP analysis is a rough-cut capacity planning (RCCP) methodology similar to many different engineering disciplines. The objective of this process is to establish the potential problems, before the crowds move through the site and are exposed to potential risks. As with all risk analysis techniques, it is not enough to document the potential risks, it is essential to implement a risk management strategy to manage or mitigate those risks. We do that by applying a value of 82 people per metre per minute, which is the maximum sustainable crowd flow at optimal density (this is a low density, less than three people per square metre). The crowds may also arrive at different times for different events, such as a concert may experience early arrivals and a fireworks display or football match may experience last minute arrivals. Which brings us to the next part of the analysis, what type of crowds are we expecting at our event?


We need to ascertain the demographic mix of the crowd and if there is any history of adverse behaviour for this location, crowd, performer or environment. There are a number of crowd classifications and crowd psychologists[1] have divided opinions on the various classifications. For crowd safety purposes we need to differentiate between the following types, each will have specific considerations for crowd safety management. The classifications we use are defined below:

Casual – People coming and going; not organized but may be in loose groups. Casual crowds will accept direction from authority and aregenerally well behaved.

 Cohesive – Crowd assembled for a specific purpose or reason such as an event concert or performance. No leadership.

 Expressive – Crowd Gathering for a common purpose. Under loose leadership or following a particular motive. Not aggressive but sections of the crowd behaviour becoming mildly anti-social. Expressive crowds may require active involvement by authorities.


Crowds engaged in acts of civil disobedience or direct action. Some sections may become aggressive and violent while other sections continue with different activities.


Crowd retreating from or reacting to a dangerous situation. This can be caused by serious anti-social behaviour and/or emergencies.

Using the above, evaluate the potential problems that may ben known in advance, of your crowd, in your space. If you have different classifications, make sure you have documented how these may affect your risk assessment and if there are specific management requirements (such as additional police resources).

The way forward

If you do nothing more than multiple the area by 2 (2 people per square metre) and evaluate egress based on exit capacity (total width of available exits multiplied by the optimal, low density, maximum crowd flow – 82 people per metre per minute) then you are doing nothing more than the minimum required for a risk analysis. You can expose the crowds to risks and you may be liable for any resulting harm caused to the crowd that can be traced back to negligent planning or approvals.

Why is this important?

The Hillsborough disaster occurred on 15 April 1989 at the Hillsborough Stadium in Sheffield, England. During the FA Cup semi-final match between Liverpool and Nottingham Forest football clubs, a human crush resulted in the deaths of 96 people and injuries to 766 others. As stated in the Hillsborough Report[2]:

It is not enough to aim only at the minimum measures necessary for safety. That has been, at best, the approach in the past and too often not even that standard has been achieved.

We have outlined a methodology that can apply to any place of public assembly. If you have responsibilities for organise, planning, managing or licensing crowded spaces. You may be interested in the Diploma in Crowd Science – Crowd Risk Analysis and Crowd Safety (Level 5). Contact us for further information ( – details are available on the website –

Just remember the crowd safety catch phrase.

Never economise on information or forethought.


About the Author

Professor Dr. G. Keith Still FIMA FICPEM SFIIRSM is the Professor of Crowd Science (Crowd Risk Analysis and Crowd Safety) at Manchester Metropolitan University. He has spent the last 25 years working in the field of crowdrisk analysis and over 15 years teaching the principles and applications of crowd science. Keith develops and runs crowd safety-training workshops around the world. His book “Introduction to Crowd Science” outlines the development and application of crowd risk analysis for built and complex spaces. Keith is a Fellow of the Institute of Mathematics and its Applications, a Fellow of the Institute of Civil Protection and Emergency Management and a Specialist Fellow of the International Institute of Risk and Safety Management. He has also consulted on some of the world’s largest events, for example, the Jamarat Bridge, annual pilgrimage to Makkah (Saudi Arabia 2001 – 2005), the Royal Wedding (UK, 2011) and Olympic events  (Sydney 2000 and London 2012).