I needed to update my model through April 20. I have not changed the model parameters for two weeks, and I only had slight changes before. Thus, the model is stable and predictive, given the faulty, incomplete data at hand. The model points out that the reported number of deaths around the world is lagging. For the world, my cautious estimate is 30,000 missing COVID-19 deaths in late March and April, see Figures 1 and 2 . My simple model points out clearly that the more complex models may not be predictive, while fitting the erratic data better. My friend ecologist, Professor Bill Rees, just reminded me that his mentor, Crawford "Buzz" Holling, insisted on repeating to his teams whenever they got bogged down in disputes over accuracy vs precision: "It is better to be approximately correct than very precisely wrong." Figure 1. My Gaussian model for the worldwide COVID-19 deaths as of April 20, 2020. The difference between the magenta extrapolation of what th...
In this blog, I continue to write about the environment, ecology, energy, complexity, and humans. Of particular interest to me are human self-delusions and mad stampedes to nowhere.