Left unchecked, the novel coronavirus (covid-19) will continue to sweep the globe. The horror stories from emergency departments and initial estimates of fatality rates starkly portend large numbers of people dying from both the virus and insufficient healthcare capacity. However, there is disagreement about what it will take to halt this progression and how far our preventative measures should go.

Our analysis in early February of 458 confirmed covid-19 cases across 93 Chinese cities was one of the clear early warnings that without interventions the average time between successive cases in a transmission chain is less than a week (around four days). Moreover, people can spread covid-19 before they even know they are sick, and there are individuals known as “super-spreaders” who infect an unusually large number of others.

These data help us understand why covid-19 is measurably more difficult to contain than a similar virus, SARS (severe acute respiratory syndrome). It comes down to a race between humans and the virus—how rapidly we can track down an infected person’s contacts versus how quickly the viral infection can incubate and spread.

While covid-19 moves quickly and sometimes silently, SARS is slow and visible. The serial interval of each demonstrates why covid-19 is the more insidious threat: if person A infects person B, the serial interval is the time between person A developing symptoms and person B developing symptoms. SARS has a serial interval of around eight days—twice as long as covid-19—and no pre-symptomatic transmission.

Imagine that a patient is diagnosed with SARS two days after first feeling sick. Public health authorities would have time on their side, perhaps a full six days to track down and isolate anyone who had contact with the patient over the prior two days. With covid-19, by the time of diagnosis a patient may have already been contagious for several days. During this period they may have infected many others who are also now spreading the virus—possibly without displaying any symptoms.

This is why covid-19 requires more drastic measures than SARS. It can spread quickly and silently, and we cannot possibly identify every infectious individual in an emerging outbreak, particularly with limited laboratory-testing capacity.

Super-spreading events compound the challenge. Among the 458 Chinese cases, five people infected over five others, with the biggest super-spreader infecting at least 16 people. Network theory—the mathematics that underlies the computer models that predict how the disease will spread— tells us that even a few people capable of infecting large numbers of others can dramatically amplify transmission and undermine interventions.

The recent threats of SARS, swine flu, Ebola, and Zika have brought fame to an epidemiological statistic known as R₀. It stands for the basic reproduction number and is intended to be an indicator of the contagiousness of infectious agents (it is pronounced R-naught). In short it tells us how many people each new case will infect during the early days of a pandemic on average. An outbreak is expected to continue if R₀ has a value >1 and to end if R₀ is <1.

A lot of attention has been paid to recent estimates suggesting that covid-19 has a lower R₀ than SARS, roughly two versus three. Clearly, then, R₀ is not the whole story. It indicates whether one case will turn into two or three or four, but not how quickly or how silently that will come to pass.

The level of intervention required to curb an outbreak very much depends on all three factors: its R₀ value, speed, and visibility in the community. We should not be fooled by the relatively modest R₀ of covid-19 as its speed and stealth make it all the more difficult to contain. Even if each case infects only two others, the number of infections can skyrocket undetected in the absence of early and extensive control measures that limit person-to-person contact.

Our study highlights the elusiveness of covid-19. Keeping people apart is the only guaranteed way to block infections given the immense challenge of identifying contagious and soon-to-be contagious cases. Whether the policy goal is to stop transmission, protect those at high risk, or "flatten the curve" to ensure that fewer people are sick at any one time, extreme social distancing strategies of the type we have been seeing are strongly recommended.

Professor Lauren Ancel Meyers is the Cooley Centennial Professor of biology and statistics at the University of Texas at Austin, where she develops powerful mathematical methods for forecasting the spread of diseases and designing effective disease control strategies.

The views and opinions expressed in this article are those of the author(s) and do not necessarily reflect the views of The Economist Group or any of its affiliates. The Economist Group cannot accept any responsibility or liability for reliance by any person on this article or any of the information, opinions or conclusions set out in the article.