Software engineer rips apart COVID-19 propagation model that locked down the world

So now we learn that the software used for modeling the propagation of COVID-19 across the world is shit software. It’s something that is over a decade old and written using poor programming practices. It appears to have been translated from one language to another, the same inputs produce different outputs (which it’s not supposed to do), and the modelers knew of all the bugs in their system and tried to cover it up by saying that averaging out the results would be accurate.

A software engineer, formerly of Google, went through the code and was appalled at what he/she found.

The Imperial code doesn’t seem to have working regression tests. They tried, but the extent of the random behaviour in their code left them defeated. On 4th April they said:  “However, we haven’t had the time to work out a scalable and maintainable way of running the regression test in a way that allows a small amount of variation, but doesn’t let the figures drift over time.”

Beyond the apparently unsalvageable nature of this specific codebase, testing model predictions faces a fundamental problem, in that the authors don’t know what the “correct” answer is until long after the fact, and by then the code has changed again anyway, thus changing the set of bugs in it. So it’s unclear what regression tests really mean for models like this – even if they had some that worked.

Dr. Fauci and Dr. Birx used this information to help convince President Trump to shut down the country. President Trump’s initial instinct that this was not that big a deal was correct and he should have stuck to his guns.

In essence, we shut down our country and the world for a virus where around 97.5% of the people that get infected fully recover. 2.5% become critical and a percentage of them die. I wish #PeakStupid wasn’t true.