You're right. I chose the optimistic end of fatality estimates because that's the conservative choice (for making the point I was making) but given that infrastructure will collapse in this scenario the true fatality rate is almost certain to be a lot higher.