I have been struggling with this question.

Let's say we have $m$ decks. One deck contains all (52) red cards. The rest of the decks contain 26 red and 26 black cards.

Given that I draw a card and it's red, what is the probability it was drawn from the biased deck (containing all red cards)?

My approach: apply Bayes' rule.

$P(d=b| c=r) = \frac{P(c=r|d=b) P(d=b)}{\sum_i(P(c=r|d_i) P(d_i)}$

But, my answer is in terms of $m$. Is that correct?

  • $\begingroup$ Think for a moment about the qualitative difference between the situation where $m=1$ and $m=1\,000\,000$. Doesn't that convince you the answer must depend on $m$? $\endgroup$ – whuber Oct 21 '16 at 19:45
  • $\begingroup$ It should depend on $m$. Clearly as $m$ gets larger the probability should tend to zero, for instance. $\endgroup$ – dsaxton Oct 21 '16 at 19:48
  • $\begingroup$ Are we assuming that the deck is selected randomly (uniformly) from the set of all decks (this seems to be assumed, but should be mentioned). $\endgroup$ – Juho Kokkala Oct 22 '16 at 6:38

You are exactly correct! Imagine if you had a million decks, the chance that it came from the biased one (1/million) is tiny compared to the 1/2*(1-1/million) $\approx$ 1/2. The total red cards in the others dominate.

Conversely if you only had 1 deck, it would be 100%!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.