That you miss the cases by half. Was listening to a podcast about the inaccuracies of the model. A 50% miss is not a 50% miss in number of cases. A 50% miss in spread rate by the end of 7 days is a 98% miss in cases. 30% miss in spread rate is 88% cases miss by 7 days and 10% miss by 7 days is 47% overstatement of cases. It’s not mitigation but that these things are so very error prone. It’s why testing and not responding to bogus models is going to be the way out of this.
