Alcohol deaths may have a high mean but lower variance, covid deaths have an unknown mean and potentially a much higher variance. So we have to prevent it from blowing up before seeing proof it could.