I'm seeing a bunch of talk about how the code behind the simulations for the March 16 @MRC_Outbreak report (imperial.ac.uk/mrc-global-inf…) was never previously released and is not well written. This is then taken to somehow mean that the "lockdown" policy was ill-founded. 1/11
However, the code and simulation mean little in terms of mortality expectations. As is usually the case, everything is about the assumptions of the model. Here, it was assumed that the fundamental reproductive number (R0) is 2.4 and infection fatality rate (IFR) is 0.9%. 2/11
In 1927(!), Kermack and McKendrick derived the total proportion of the population expected to be infected in the course of an epidemic in a simple SIR model as the attached equation. Solving this for R0 of 2.4, we get a total proportion infected (or attack rate) of 88%. 3/11 
From the report, running the simulation gives an attack rate of 81%. As you'd expect, particular choices about household size, etc... make a difference, but it's not huge. 4/11
Without the simulation, just from basic epi knowledge we can assume that COVID-19 will infect a large fraction of the population in the absence of any control measures or *behavioral changes*. 5/11
Given an IFR of 0.9% and ~80% of the population infected, we easily derive the widely circulated mortality estimates of ~2M deaths in the US and ~500k deaths in the UK. 6/11
I'm all for transparency in scientific software and have pushed for reproducible computational research for years, but the code behind the @MRC_Outbreak simulation is not the fundamental issue here. 7/11
Key questions surround the relative effectiveness of lockdown policies vs natural behavioral changes in reducing R0, but not what would happen in a hypothetical "do nothing" scenario. 8/11
The biggest issue with the "do nothing" scenario of ~2M deaths is that it would never come to pass due to natural behavioral changes. 9/11
However, the main point of the report was that given IFR, we should be pursuing suppression rather than mitigation. This implies a strict lockdown for suppression followed by #TestTraceIsolate to keep epidemic suppressed. 10/11
Notably, this is exactly what countries like South Korea and New Zealand have been able to achieve. The US was not able to reach suppression with our lockdown and so we're left with agonizing decisions about how to keep society functioning while holding the virus in check. 11/11
