is why you are seeing the massive change in the predictions. I don't know what the change to the model is but it did cause those huge shifts that you mentioned in your post. I'm guessing it ends up being more accurate that what they were predicting earlier, but the question is how accurate. At least they were willing to recognize their model had some serious flaws to it and were willing to make a significant adjustment to it to hopefully improve it (especially since this model seems to be on of the primary ones used by the decision makers in Washington).

Part of the problem with this virus is there is/has been so much unknown about it. All these models have been built all based upon unknown information and a lot of them are producing very different results. As more becomes known and the models shift with the additional information I like to think they are becoming more accurate and are starting to converge on what we should actually expect, but that's something that we unfortunately probably won't know until after this thing plays out.