I guess what I'm saying is that the reason the mortality rate could be lower is because you can't really get an accurate % rate unless you have an idea of how many people truly have it, otherwise the date will be skewed... I mean, 38 deaths out of 1,500 confirmed cases for example would give...