We conducted 250 thousand simulations of the global progress of the COVID-10 outbreak with the intensity of a seasonal reduction in transmission sampled from a flat distribution from 0 to 100%. Other parameter values were sampled according to approximations of the distribution of values described in the literature. The simulation starts on the 5th of February and each simulation is selected according to its fit to observed data up to the 18th of February.
The fit was scored using a least squares sum of the differences compared to the data available so far. The top 1,000 best fitting scenarios were selected from the 250,000 to show the distribution of outcomes here.
We also included response scenarios that reduce the transmissibility over time to a minimum value (between 0.25 and 0.75) that is sustained indefinitely and a decay rate parameter between 0 and 1.