1

This question is not about how to solve the problem, but is about why doesn't my solution work.

A bowl has $2$ red, $2$ green, and $2$ blue marbles. How many combinations are possible if we take $3$ random marbles at a time?

I know that the answer should be, $7$.

Coz,

  1. BBG
  2. BBR
  3. BGG
  4. BGR
  5. BRR
  6. GGR
  7. GRR

But, when I want to calculate this using a combination formula, nothing comes out which is near $7$:

$\frac{^6C_3}{2! \cdot 2! \cdot 2!}$

$ = \frac{6!}{3! \cdot (6-3)!} \cdot \frac{1}{2! \cdot 2! \cdot 2!}$

$ = \frac{6!}{3! \cdot 3!} \cdot \frac{1}{2! \cdot 2! \cdot 2!}$

$ = \frac{6 \cdot 5 \cdot 4 \cdot 3 \cdot 2 \cdot 1}{3 \cdot 2 \cdot 1 \cdot 3 \cdot 2 \cdot 1} \cdot \frac{1}{2! \cdot 2! \cdot 2!}$

$ = \frac{6 \cdot 5 \cdot 4 }{3 \cdot 2 \cdot 1} \cdot \frac{1}{2! \cdot 2! \cdot 2!}$

$ = \frac{5 \cdot 4 }{2! \cdot 2! \cdot 2!}$

$ = \frac{5}{2!} $

$ = \frac{5}{2}$

So, what am I doing wrong?

user366312
  • 1,641
  • Could you explain how you found that $\frac{5}{2}$ value? – Bill O'Haran May 03 '18 at 17:00
  • So... the numerator seems to be described by taking three of six distinguishable objects and the denominator seems to be described by "forgetting" which object of which type was which. An example of a correct time to use the denominator would be with $\frac{6!}{2!2!2!}$, the number of ways in which you can lay all six of them out and order mattering. This is not useful for your current problem however since we are only taking three of them out (and the remaining three can be in whatever order, we don't care). – JMoravitz May 03 '18 at 17:05
  • As for a correct method to arrive at an answer, I would temporarily ignore that there are only two of each color available, count how many arrangements there are without that restriction, and then remove the invalid arrangements. (There are many more than $10$ if you consider order within the arrangement to be relevant and many fewer than $10$ if you don't, so you counted incorrectly by hand). – JMoravitz May 03 '18 at 17:07
  • 1
    Also, I do not get how BRB and BBR (6 and 2) are different if the marbles are picked at once. – Bill O'Haran May 03 '18 at 17:08
  • Why do you imply combinations BBG and BGB to be different? And if so, why don't you consider combinations RGB for example? – D F May 03 '18 at 17:09
  • "Could you explain how you found that $\frac{5}{2}$ value" that BillO'Haran said earlier wasn't in reference to how to continue the arithmetic in order to simplify $\frac{\binom{6}{3}}{(2!)^3}$ into $\frac{5}{2}$, it was a question as to why you would think to divide $\binom{6}{3}$ by $(2!)^3$ in the first place – JMoravitz May 03 '18 at 17:10
  • Okay. Just making sure :) – Bill O'Haran May 03 '18 at 17:10
  • If order doesn't matter, then you still have BRG and BGR both appearing on your list. This is just another reminder as to why brute force counting is a bad idea as it is very easy to make mistakes like this. – JMoravitz May 03 '18 at 17:14

3 Answers3

3

You can essentially break the possibilities down into 3 cases:

  • all the colors are the same; that cannot happen so it yields 0 possibilities

  • two marbles have the same color and the last one is different, as much possibilities as ordered couples of colors (6 possibilities).

  • all marbles are different, 1 possibility

In the end, you get 7 possibilities.

Bill O'Haran
  • 3,034
  • 1
    @yahoo.com Don't. Or at least, understand where and why the formula is used. It is much better to use techniques than formulas. – JMoravitz May 03 '18 at 17:13
  • I agree with JMoravitz. Combinatronics is about counting so adapting is key especially with small fixed numbers. However, if this exercise was meant to be generalized, then it takes it to a whole new level. Still, notice that breaking your problem down into cases is a tool just as useful as combinations and permutations. – Bill O'Haran May 03 '18 at 17:18
  • The techniques used here are the rule of sum and the rule of product, nothing more and nothing less. The summation appears in recognizing you can add the results of the three individual cases. The product appears when counting the number of arrangements in case2 by first picking which color appeared twice followed by which color appeared once ($3\cdot 2=6$). If you wish to generalize the problem, you could do the same, or you may take shortcuts using inclusion-exclusion and stars-and-bars. – JMoravitz May 03 '18 at 17:18
  • @yahoo.com see my answer below to the question of "why doesn't mny solution work." As for your claim that the question is not in fact "how to solve the problem" I point you to your choice of title: How can I solve this Combination with indistinguishable-objects problem? – JMoravitz May 03 '18 at 17:48
2

"The question is not about how to solve the problem, but was about why doesn't my solution work."

Continuing from an earlier point I made in the comments, it is much more important to rely on techniques than on formulas at this early level. Any formula you encounter can be replicated on the spot whenever you need it using the techniques you have learned, including combinations, permutations, stars and bars, combinations with repitition, etc...

I mentioned elsewhere how rule of sum and rule of product are essential introductory counting techniques.

Another very useful counting technique is sometimes referred to as the following:

The Shepherd's Principle: When counting objects or arrangements, if your current count is $n$ but you have overcounted each object/arrangement the exact same number of times, say $k$ times each, then you may correct your count by dividing: $\frac{n}{k}$

The name comes from that if you know that every sheep has four legs each and you wish to count the number of sheep in your field and can see every leg but cannot distinguish one body from another or just don't have a clear view of the bodies, you may count how many legs you see and divide by four.

The shepherd's principle explains where most of the division comes from in formulas and techniques used in counting. It can be seen for example in the formula for combinations, $\binom{n}{r}=\frac{n!}{r!(n-r)!}$, and is one among many ways of proving that formula to be correct. Related to your example, the number of ways to arrange all six of the marbles in a line where order matters would be $\frac{6!}{2!2!2!}$, recognized by temporarily assuming the balls are all different colors making $6!$ arrangements and recognizing that each arrangement was overcounted a total of $2!2!2!$ times each.

Why didn't it work for you here? Because many of the arrangements you were counting were not counted $2!2!2!$ times each and some arrangements were overcounted more times than others.

Among the $\binom{6}{3}=20$ ways in which we may select three marbles if we were to temporarily assume they were all distinct, it breaks into cases. Eight of which correspond to all of the colors actually being different. The remaining twelve ways correspond to two of the colors actually matching and the remaining color being different. In the case that all three balls were actually different colors, this was accidentally counted $2!2!2!=8$ times but should only have been counted once. In the case that two of the balls were the same actual color, these were only accidentally counted twice each, not eight times each, yielding a corrected count for this case as being $\frac{12}{2}=6$.

This would give the corrected count as $\frac{8}{2!2!2!}+\frac{12}{2}=1+6=7$, giving the correct answer.

I personally find this to be a rather unhelpful technique to be used for this problem, but it can be forced to work, especially since in order to have counted the number of arrangements which fell into the specific cases it requires as much effort as going through the problem a more direct way to begin with essentially making the application of the shepherd's principle unnecessary extra work.

JMoravitz
  • 79,518
0

use generating function to count :

you can select either $0$ ball ; $1$ ball ; $2$ ball of same colour

so $f(x)= (x^0+x+x^2)(x^0+x+x^2)(x^0+x+x^2)=(1+x+x^2)^3$

so, to find number of ways to pick $3$ balls out of the bowl, just find coefficient of $x^3$ in expansion of $f(x)$

now , coefficient of $x^3$ in expansion of $f(x)=(1+x)^3+3.(1+x)^2.x^2+3(1+x)x^4+x^6$ is $7$ so there will be 7 combinations of drawing a ball from the bowl.

  • @ mr yahoo: your solution is completaly wrong ....4 minutes earlier you came up with 10 cases and now you are saying "i know the answer is 7" and you just edited your question .......it seems like you know yourself where you were wrong –  May 03 '18 at 17:27
  • @ yahoo: in your numerator you used , the combination formula...but $^{n}C_{r}$ is only applied when selecting$ r$ different things out of $n$ different things....but here balls of same colour are identical / indistinguishable so $^{n}C_{r}$ will not work here –  May 03 '18 at 17:33
  • 1
    @ mr. veeresh pandey : 4 minutes earlier you came up with 10 cases and now you are saying "i know the answer is 7" and you just edited your question --- read the comments and you will know why. .......it seems like you know yourself where you were wrong --- yeah, nice guess! – user366312 May 03 '18 at 18:12