COVID-19 Day 75: Checking Back on Singapore’s Pandemic Predictions

At least they fixed this for the Sonic movie. The Imperial College and SUTD, not so much.

I discovered that The Singapore University of Technology and Design (SUTD) had hidden their SIR model, which showed their predictions for when each country’s COVID-19 pandemic would end. In their words, “Over-optimism based on some predictions is dangerous because it may loosen our disciplines and controls and cause the turnaround of the virus and infection, and must be avoided. Earlier predictions are no longer valid because the real-world scenarios have changed rapidly.”

I recently rediscovered the old SUTD predictive models, in of all places, my computer. I had forgotten that I downloaded their paper (which was available for download on their page) weeks ago. In the archive folder were their charts as of April 26, 2020. Now, a month later, I thought it would be worthwhile to see how well their predictions turned out for the UK, Sweden, Switzerland, and the USA.

UNITED KINGDOM

© SUTD SIR model

Johns Hopkin’s COVID Tracker

SWEDEN

© SUTD SIR model
Johns Hopkin’s COVID Tracker

SWITZERLAND

© SUTD SIR model
Johns Hopkin’s COVID Tracker

USA

© SUTD SIR model
Johns Hopkin’s COVID Tracker

As we can see with 20/20 hindsight, the SUTD model was only close for Switzerland predicting a 99% probability that their outbreak would end by May 8. Today Switzerland has brought their daily new cases down to just 30, not quite an end but very close. It appears their COVID-19 mitigations efforts are as efficient and predictable as their trains.

The SUTD model predicted the end of COVID in the US would occur around May 32 but that day came and went with still around 20,000 new cases per day as of this writing. The US needs to get its daily new cases down by a factor of 12 or more to approach Switzerland’s level of control. The overall trend is downward but looking less and less like the SUTD’s bell curve and more like the slow decline of a bacterial growth curve.

Unfortunately for the US, UK, and Sweden, their cases have been declining but not nearly at the rate predicted by the SUTD model. Sweden in particular, has flattened their daily cases to around 500, perhaps truly trying for herd-immunity? Only time will tell if the summer’s sun and warm weather will aid in reducing numbers. Or if COVID-fatigue will set and people relax their safety measures bringing about a spike in new cases for the rest of Europe and North America.

As with the infamous Imperial College paper, predictive models aren’t predictions of the future, but tools to determine probable outcomes based on the data and assumptions put into them. But the data and the ground-truth can change, such as more testing uncovering more cases, weather, political action, or reopening of public businesses. When the variables change, all bets are off.

SOURCES

https://coronavirus.jhu.edu/data/new-cases

https://www.thoughtco.com/bacterial-growth-curve-phases-4172692

Comments are closed.