Since the cat is out of the bag, then compare the model with real world data. A little too late to be useful if what I am hearing about the models is correct.
The real world data reflects the lock down.
You don’t have much real world data that doesn’t reflect a lock down.
You can’t compare a model with no lock down with a model that has a lock down.
And HCQ has been a game changer too. The models didn’t predict that.
So now you have data that doesn’t match the model, but you have changes that the model didn’t consider. Thus you can’t say the model was wrong.