Probability all balls are white.

  • Thread starter Thread starter Swetasuria
  • Start date Start date
  • Tags Tags
    Balls Probability
AI Thread Summary
The discussion revolves around calculating the probability that all balls in a bag are white after drawing two white balls. Initially, five scenarios for the composition of the balls are considered, leading to an initial probability of 1/5 for all being white. However, after drawing two white balls, the probability must be recalculated using Bayes' theorem, which incorporates the likelihood of each scenario given the new information. The final calculation indicates that the probability of all balls being white is actually 3/5, highlighting the importance of prior probabilities in Bayesian analysis. The conversation emphasizes the need for a priori assumptions about the distribution of colors in the bag to accurately determine probabilities.
Swetasuria
Messages
48
Reaction score
0

Homework Statement


A bag contains 4 balls. Two balls are drawn at random, and are found to be white. What is the probability that all balls are white?

2. Relevent equations
At school, I'm currently learning Bayes' theorem, probability disribution and Bernoulli trials.

The Attempt at a Solution


The 4 balls can be
1) all white
2) 3 white, 1 non-white
3) 2 white, 2 non-white
4) 1 white, 3 non-white
5) all non-white

P(all white)=1/5

But given, two balls taken at random are white.

I can't figure out what to do next. When I ignore the 2 white balls, I get the answer as 1/3 which is wrong according to my textbook.

[Edit]: I got it!
A-2 balls taken are white
E1-4W
E2-3W, 1Non-white
E3-2W, 2N
E4-1W, 3N
E5-4N

P(E1/A)= {P(E1)P(A/E1)} / {P(E1)P(A/E1)+...+P(E5)P(A/E5)}

={1/5*1} / {1/5*1 + 1/5*3C2/4C2 + 1/5*2C2/4C2 + 1/5*0 + 1/5*0}
=3/5
 
Last edited:
Physics news on Phys.org
Swetasuria said:
P(all white)=1/5
Does the problem statement say that all 5 cases have the same probability?
Otherwise, you cannot know this.
 
I'm not too good at this, so may have it wrong, but it seems to me that once you have pulled 2 balls and they are white, what you have is a situation where you have 2 balls (remaining) and you want to know what are the odds that they are both white.

This assumes that the starting condition becomes irrelevant once you draw the first two, and that's where I may be going wrong, though I don't think so.
 
As Bayes showed, the problem is not solvable without plugging in some a priori probabilities. E.g. you could argue that, by and large, bags of balls tended to be, about equally often say, mixed colours or all the same colour. This would be for reasons related to why balls are put in bags in the first place.
Another reasonable choice says each ball is independently white or non-white, but since there are many possible colours these might not be with prob 1/2. Let's say a ball is white with prob p. Now we find that drawing two white balls tells us nothing about the remaining two (as phinds surmised). The prob that they are both white is p2.
 
Back
Top