The latest Spotty ™ winner:
What is the difference between a forecast model and a simulation model?
For his letter in today’s Star Tribune (5-3-20), Richard Holman wins a Spotty™.
Why are the model outputs so different? The April 24 article “Why Minnesota’s COVID-19 models are so different” did a good job of explaining. What it should have added is that the Minnesota model is a simulation model and the Washington model is a forecast model.
Forecast models use past patterns to predict future patterns. Most businesses that project future sales use a forecast model. These models can be quite accurate if the future environment or conditions are similar to the past. Therefore, until several countries have a major second wave of COVID-19 infections, the Washington model will not be predicting a second wave.
Simulation models use past data, but not necessarily past patterns, to assess what might happen if future conditions are different from the past, or there is no history of the past. Manufacturers use simulation models to assess what will happen to production throughput if they make major changes to a production line. These models are very good for “what if” analysis.
So which model will be more accurate for COVID-19 projections? If the past is a good indicator of the future, the Washington forecast model will be. But, if future conditions are significantly different, for instance rapidly moving away from social distancing, the Minnesota simulation model will be the better model to predict the outcome.
Richard Holman, Loretto
What Mr. Holman is saying, perhaps, is that one model or the other will prove more accurate depending on how much of this kind of stuff we see in coming days.
I wrote a limerick about this.
Consider the loudmouth, the JLew,
His appeal to some voters, the yahoo,
He rubs shoulders with them,
They don’t know where he’s been,
It’s a corona convention impromptu.
Remember, a Spotty™ is awarded for a letter to the editor, an op-ed, or a blog post or comment that I wish I had written myself.
Thanks for your feedback. If we like what you have to say, it may appear in a future post of reader reactions.