In a certain day care class, $30\%$ of the children have grey eyes, $50\%$ of them have blue and the other $20\%$'s eyes are in other colors. One day they play a game together. In the first run, $65\%$ of the grey eye ones, $82\%$ of the blue eyed ones and $50\%$ of the children with other eye color were selected. Now, if a child is selected randomly from the class, and we know that he/she was not in the first game, what is the probability that the child has blue eyes?

My solution

Let's say $B =$ blue, $G =$ grey and $O =$ "Other color" and $NR =$ "not selected for the first run"

$$P(B \mid NR) = \frac{P(NR \mid B)P(B)}{P(G)P(NR \mid G) + P(B)P(NR \mid B) + P(O)P(NR \mid O)}$$

On substituting values $$P(B \mid NR) = \frac{0.5 \cdot (1-0.82)}{(0.3 \cdot (1-0.65)) + (0.5 \cdot (1-0.82)) + (0.2 \cdot (1-0.5))}$$

$$P(B \mid NR) = 0.305$$

Is this the right way to use bayes theorem?


Yes, this is correct.

The general form of Bayes' rule is

$$P(A_i|B)= \frac{P(B|A_i)P(A_i)}{\sum_jP(B|A_j)P(A_j)}$$

as you've used above.

| cite | improve this answer | |
  • $\begingroup$ Thanks for the generalized version of Bayes' theorem. Do you see anything wrong with the way I have arrived at the individual probabilities? $\endgroup$ – naresh Nov 16 '12 at 17:12
  • $\begingroup$ Nothing wrong that I can see... $\endgroup$ – Simon Hayward Nov 16 '12 at 18:49
  • $\begingroup$ Has that helped, do you have the answer now? $\endgroup$ – Simon Hayward Nov 16 '12 at 19:23

Suppose 100x total students

Blue eyed =50x

Grey eyed =30x

Other eyed =20x

1st run selection :

Blue eyed =41x

Grey eyed =19.5x

Other eyed=10x

Not selected in 1st round

Blue= 9x


Other =10x


| cite | improve this answer | |

$$p(\textrm{not game 1} \mid \textrm{blue}) \div p(\textrm{not game 1}) = 0.18×0.50 \div (0.18×0.50+0.5×0.2+0.35×0.30) =0.3051$$

| cite | improve this answer | |

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.