Just taking a wild shot at this one, but I suspect that the mistake is between C and D. In C, you start with an even distribution over all the people in the experiment, and then condition on surviving. In D, your uncertainty gets allocated among the people who have survived the experiment. Once you know the rules, in C, the filter is in your future, and in D, the filter is in your past.
Just taking a wild shot at this one, but I suspect that the mistake is between C and D. In C, you start with an even distribution over all the people in the experiment, and then condition on surviving. In D, your uncertainty gets allocated among the people who have survived the experiment. Once you know the rules, in C, the filter is in your future, and in D, the filter is in your past.