Daniel K. answered 11/17/21
Masters in Statistics for APs, AMC training, Calc-based Statistics, R
(ii) The simplest way to do this is to get 1 - P(less than two orange) = 1 - P(no orange) - P(exactly one orange), so that's 1 - (P(green ball in one draw))^8 - (8 choose 1)*((P(one green in one draw))^7)*(P(one orange in one draw)) = 1 - ((12/20)^8) - 8*((12/20)^7)*(8/20) = .8936
(iii) The only way to have equal numbers of green and orange is to have exactly four of each. That's (8 choose 4)*((P(one green in one draw))^4)*((P(one orange in one draw))^4) = 70*((12/20)^4)*((8/20)^4) = .2322
(b) We need the probability of getting at least one green ball to be at least .99 . Because that situation includes many situations, I would rather choose to deal with the opposite. 1 - P(zero green balls) >= .99 which means P(zero green balls) <= .01 . Now I need to express P(zero green balls in k draws) based on k. This is simply P(not a green ball in one draw)^k = (8/20)^k so I need (8/20)^k <= .01 . If I don't want to guess and check I use logarithms, log((8/20)^k) <= log(.01) and so k*log((8/20)) <= log(.01) which gives k >= log(.01)/log((8/20)) . (Note that the inequality switches because log((8/20)) is negative. Multiplying or dividing by a negative switches the direction of the inequality.) So we have k >= 5.0259 . Since k must be an integer we need to choose 6. We need to choose at least 6 balls to get a 99% probability of having at least one green.
Aamir H.
Thank you11/18/21