Menu
Log in


Dice Killer Paradox

  • 21 Apr 2025 4:06 PM
    Reply # 13489951 on 13486970

    Hi Chris,

    I'm not familiar with the paradox - thanks for introducing it! I think that the argument that you should toss the coin is wrong because even though more than half of the victims die, the probability that the cohort that you belong to is among the dead is 1/6.  Intuitively we want to say that the fact that more than half of the victims will die means that we have more than a half probability of dying - but that's not true.  The victims are not exchangeable, because of the stopping rule.  

    At least, that's what I think for now. ;-)

    Cheers

    Andrew

  • 14 Apr 2025 5:02 PM
    Message # 13486970
    The Dice Killer Paradox


    This has the strong whiff of Monty Hall. But is not the same because Monty cannot choose the door to reveal to you.

    Do members know of this paradox? If so, can they enlighten me as to how the rather daft argument that you should toss a coin is wrong? Otherwise, I will have another sleepless night!

    One argument that occurs to me is that if you were told what round you were in, then the justification for the coin toss evaporates. You are not 1 member of an increasing unknown sequence. You will take the roll of the dice.

    The round you get to is completely random and a appears irrelevant to decision making. My PhD was on conditional inference and so-called ancillary statistics and the number of rolls of a die without hitting 6 looks pretty damned ancillary to me! But the catch is that it is unobserved so we cannot condition on it.


    Last modified: 14 Apr 2025 5:34 PM | Chris Lloyd
Powered by Wild Apricot Membership Software