I am an expert in statistics. I spent my whole life writing statistical software and running analysis. I came up with numerous statistical models on human behavior and I can tell you this. The models can say whatever you want them to say. Tweak the data a very small bit and your result is twice as big or twice as small. Math models are very easy to make and very hard to make right. A good example may be the weather 14 days out. Hundreds of millions of dollars are spent on those prediction models for weather. But they are no better than a monkey throwing darts.
When someone says that their model predicts a million deaths, its because the makers of the model want it to predict a million deaths. They could have made it show fifty thousand deaths but they did not want it. A model is like an expert, you can find one that will say anything.
poing,
Thank you very much for the truth about statistical models.
I’m educated in statistics too, but not to your level. I agree with your statements. I noticed something interesting at the worldometer cv site. Go to Colorado and hit the “projections” link on the far right.
Scroll down to daily deaths. You can compare Colorado’s rate of daily deaths to the other states on a graph. I was randomly doing this and was surprised at how often the smaller waves of rise and fall within the entire curve of Colorado would line up with different states’ small rises and falls within their larger curves. This happens even with some states that have much more severe curves and much larger case loads.
I’m not drawing any conclusions but I’m going to investigate that more.