That's probably the best one because you don't have to infer the denominator. He gives a range around his estimate. Most everyone who died was over 65. The death rate for the flu for 65+ is about 1 percent. So the number of deaths is about what you might expect if the same people were infected with the flu. This is a conclusion I had come to before reading this paper. Of all the data we have, the cruise data is probably the most "pure". The main problem is that it's a small sample.