Probability Question: Please Experts, settle once & for

This topic has expert replies
User avatar
Junior | Next Rank: 30 Posts
Posts: 20
Joined: Sun Aug 08, 2010 8:36 pm
Please answer and clarify why and how to solve such a question: I will appreciate if the experts (instructors) here can definitely clarify this for posterity's sake:

There are 3 white 3 black and 2 red balls.

1. Balls are drawn one by one WITH replacement: What is the probability that the third ball drawn is RED?
2. Balls are drawn one by one WITHOUT replacement: What is the probability that the third ball drawn is RED?
3. 3 Balls are drawn at once (at one go - I assume in this case with or without replacement does not make sense, correct): What is the probability that there is one RED ball?

Thanks
Source: — Problem Solving |

GMAT/MBA Expert

User avatar
GMAT Instructor
Posts: 2623
Joined: Mon Jun 02, 2008 3:17 am
Location: Montreal
Thanked: 1090 times
Followed by:355 members
GMAT Score:780

by Ian Stewart » Tue Aug 10, 2010 10:32 pm
mainhoon wrote:Please answer and clarify why and how to solve such a question: I will appreciate if the experts (instructors) here can definitely clarify this for posterity's sake:

There are 3 white 3 black and 2 red balls.

1. Balls are drawn one by one WITH replacement: What is the probability that the third ball drawn is RED?
If we are selecting with replacement, then every time we make a selection we have 3 white, 3 black and 2 red balls. So the probability of selecting a red on any selection would be 2/8.
mainhoon wrote: 2. Balls are drawn one by one WITHOUT replacement: What is the probability that the third ball drawn is RED?
The fact that it's the third ball doesn't matter. The third ball is just as likely to be red as any other ball, so the answer will again be 2/8.
mainhoon wrote: 3. 3 Balls are drawn at once (at one go - I assume in this case with or without replacement does not make sense, correct): What is the probability that there is one RED ball?
Drawing 'three at once' is mathematically the same as drawing three balls one at a time without replacement (it doesn't matter if there's a nanosecond between each selection or if they all happen exactly simultaneously), so you could look at this situation from either perspective. I assume you're asking about there being exactly one red ball.

If you imagine picking 'three at once', so order does not matter, we have 8C3 ways of picking three balls from eight, 2C1 ways of picking one red ball from two, and 6C2 ways of choosing two non-red balls from the remaining six, so the probability would be (6C2)(2C1)/(8C3).

If instead you imagine picking the balls one at a time, there are 3 ways to get exactly one red ball: we can pick it first, second or third. Each of those situations is equally probable, so we can find the probability of picking a red, non-red and non-red in order, and then multiply by 3. Thus the answer is 3*(2/8)(6/7)(5/6).

In either case, if you simplify, you'll get the answer 15/28.
For online GMAT math tutoring, or to buy my higher-level Quant books and problem sets, contact me at ianstewartgmat at gmail.com

ianstewartgmat.com

Junior | Next Rank: 30 Posts
Posts: 25
Joined: Thu Aug 05, 2010 3:43 am

by sheelanadh » Wed Aug 11, 2010 4:22 am
mainhoon wrote:Please answer and clarify why and how to solve such a question: I will appreciate if the experts (instructors) here can definitely clarify this for posterity's sake:

There are 3 white 3 black and 2 red balls.

1. Balls are drawn one by one WITH replacement: What is the probability that the third ball drawn is RED?
2. Balls are drawn one by one WITHOUT replacement: What is the probability that the third ball drawn is RED?
3. 3 Balls are drawn at once (at one go - I assume in this case with or without replacement does not make sense, correct): What is the probability that there is one RED ball?

Thanks
what is OA?

User avatar
Junior | Next Rank: 30 Posts
Posts: 20
Joined: Sun Aug 08, 2010 8:36 pm

by mainhoon » Wed Aug 11, 2010 6:42 am
sheelanadh wrote:
mainhoon wrote:Please answer and clarify why and how to solve such a question: I will appreciate if the experts (instructors) here can definitely clarify this for posterity's sake:

There are 3 white 3 black and 2 red balls.

1. Balls are drawn one by one WITH replacement: What is the probability that the third ball drawn is RED?
2. Balls are drawn one by one WITHOUT replacement: What is the probability that the third ball drawn is RED?
3. 3 Balls are drawn at once (at one go - I assume in this case with or without replacement does not make sense, correct): What is the probability that there is one RED ball?

Thanks
what is OA?
No OA. I saw this question somewhere and made up the 3 options. Ian is correct.

User avatar
Junior | Next Rank: 30 Posts
Posts: 20
Joined: Sun Aug 08, 2010 8:36 pm

by mainhoon » Wed Aug 11, 2010 6:48 am
Ian Stewart wrote:
mainhoon wrote:Please answer and clarify why and how to solve such a question: I will appreciate if the experts (instructors) here can definitely clarify this for posterity's sake:

There are 3 white 3 black and 2 red balls.

1. Balls are drawn one by one WITH replacement: What is the probability that the third ball drawn is RED?
If we are selecting with replacement, then every time we make a selection we have 3 white, 3 black and 2 red balls. So the probability of selecting a red on any selection would be 2/8.
mainhoon wrote: 2. Balls are drawn one by one WITHOUT replacement: What is the probability that the third ball drawn is RED?
The fact that it's the third ball doesn't matter. The third ball is just as likely to be red as any other ball, so the answer will again be 2/8.
mainhoon wrote: 3. 3 Balls are drawn at once (at one go - I assume in this case with or without replacement does not make sense, correct): What is the probability that there is one RED ball?
Drawing 'three at once' is mathematically the same as drawing three balls one at a time without replacement (it doesn't matter if there's a nanosecond between each selection or if they all happen exactly simultaneously), so you could look at this situation from either perspective. I assume you're asking about there being exactly one red ball.

If you imagine picking 'three at once', so order does not matter, we have 8C3 ways of picking three balls from eight, 2C1 ways of picking one red ball from two, and 6C2 ways of choosing two non-red balls from the remaining six, so the probability would be (6C2)(2C1)/(8C3).

If instead you imagine picking the balls one at a time, there are 3 ways to get exactly one red ball: we can pick it first, second or third. Each of those situations is equally probable, so we can find the probability of picking a red, non-red and non-red in order, and then multiply by 3. Thus the answer is 3*(2/8)(6/7)(5/6).

In either case, if you simplify, you'll get the answer 15/28.
Excellent Ian. I had sort of made up the options to understand the thinking process. What boggles my mind is that 1 and 2 give the same answer - I had heard that was the case but the logic was not clear to me. Let me indulge you for a second. Let's say we have 1000 Red and 1 Black ball. If I want to find the probability of finding the black ball:
1. with replacement: clear answer 1/1001
2. without replacement: lets say in the 60th try = as I draw the balls, it would seem that the probabilty of finding the black ball should improve (because the red ball population is depleting) so the larger number of tries the higher the probability of finding the black ball. However as you suggest and I believe is also correct, the probability is the same 1/1001. How do I rationalize this?
To generalize this further on such a question then, all one needs to do is to find the probability in the initial population and that is the answer regardless.

On a subtle note, I have an issue with the wording around such problems "draw n balls with replacement" - unless this is done one by one it makes no sense, so why does it not say one by one. In other words this can equally mean that I just grab n balls at once (which is my case 3 above). Is the fact that I add with replacement makes it now a 1 by 1 case?

Thanks again!

GMAT/MBA Expert

User avatar
GMAT Instructor
Posts: 2623
Joined: Mon Jun 02, 2008 3:17 am
Location: Montreal
Thanked: 1090 times
Followed by:355 members
GMAT Score:780

by Ian Stewart » Wed Aug 11, 2010 12:13 pm
mainhoon wrote: Excellent Ian. I had sort of made up the options to understand the thinking process. What boggles my mind is that 1 and 2 give the same answer - I had heard that was the case but the logic was not clear to me. Let me indulge you for a second. Let's say we have 1000 Red and 1 Black ball. If I want to find the probability of finding the black ball:
1. with replacement: clear answer 1/1001
2. without replacement: lets say in the 60th try = as I draw the balls, it would seem that the probabilty of finding the black ball should improve (because the red ball population is depleting) so the larger number of tries the higher the probability of finding the black ball.
When you pick the 61st ball in your example, there are two possible situations. Either we have not yet picked the black ball, in which case, as you say, it's a bit more likely than 1/1001 that we pick the black ball. But it can also happen that we've already picked the black ball in one of our first 60 selections, in which case there is *no* chance our 61st selection is black. These probabilities will 'balance out' to give the answer 1/1001.

It might be easier to see with a different example. If you have 1001 people, 1 of whom is a man and 1000 of whom are women, and you line them up, what's the probability the 61st person in line is the man? Well, there's nothing special about the 61st place in line; when we pick the 61st person, we are selecting from 1001 people, 1 of whom is a man, so the probability is 1/1001. The same would be true for any other position in the line.

Now, there's a different kind of question that you might be thinking of here - a conditional probability question. We might instead ask: If you have 1001 balls, one of which is black, and you pick 61 balls one at a time without replacement, if none of the first 60 selections is black what is the probability the 61st ball is black? Notice the difference between this question and the one you asked above - here we *know* that we haven't yet picked the black ball when we make our 61st selection. So we know that we have 1 black ball among the 941 balls we are selecting from (we've removed 60 red balls), so the answer is 1/941.

mainhoon wrote:
On a subtle note, I have an issue with the wording around such problems "draw n balls with replacement" - unless this is done one by one it makes no sense, so why does it not say one by one. In other words this can equally mean that I just grab n balls at once (which is my case 3 above). Is the fact that I add with replacement makes it now a 1 by 1 case?
I'm not sure I quite understand the question here, but you might look at this situation as follows: say you have a bag of marbles, and you need to pick two of them without replacement. You can put both of your hands in the bag and grab two marbles. Now, whether you take both of your hands out of the bag at once, or take your left hand out first and then your right, the results you can get will be identical. That's why it's logically equivalent to pick 'two marbles at once' or to pick one marble, then the next, without replacement.
For online GMAT math tutoring, or to buy my higher-level Quant books and problem sets, contact me at ianstewartgmat at gmail.com

ianstewartgmat.com

User avatar
Junior | Next Rank: 30 Posts
Posts: 20
Joined: Sun Aug 08, 2010 8:36 pm

by mainhoon » Wed Aug 11, 2010 12:22 pm
Yes you answered my questions perfectly. Since the with replacement and without replacement answers are identical, I wonder if you can give me an example of a probability question, say dealing with balls again, where they ARE different and is GMAT compatible - I guess there are more complicated questions out there, but wanted to know GMAT level.

Also, another question I have is this:

How many ways can I give 15 different rocks to 10 people? The answer presumably is 15^10

How many ways can I give 15 different rocks to 10 people so that each has at least 1? Is then the answer 15!/5! x 5^10

How do the answers change if

How many ways can I give 15 identical rocks to 10 people? Answer is (15+10-1)C(10-1)?

How many ways can I give 15 identical rocks to 10 people so that each has at least 1? Each gets 1, so 10 gone, 5 to go. How do I give 5 identical rocks to 10 people? 5^10?