Is This Binomial Coefficient Identity True?

In summary: Sudharaka.You are correct, thank you for pointing out the mistake. I have typed it incorrectly. It should be,\[\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\sum_{u=0}^{k-1}{k+1 \choose k-u}{n-k-1 \choose u}-\sum_{u=0}^{k-1}{k \choose k-u}{n-k-1 \choose u}\]Then, the upper bound of the second summation is $k-1$ and $u=0$ is the lower bound. The last term of the sum is,\[{
  • #1
VincentP
7
0
I'm having trouble proving the following identity (I don't even know if it's true):
$$\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\binom{n-1}{k-1}$$ $$\forall n,k \in \mathbb{N} : n>k$$
Thank you in advance for any help!
Vincent
 
Physics news on Phys.org
  • #2
VincentP said:
I'm having trouble proving the following identity (I don't even know if it's true):
$$\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\binom{n-1}{k-1}$$ $$\forall n,k \in \mathbb{N} : n>k$$
Thank you in advance for any help!
Vincent

Maybe expanding the binomial coefficients would help. Recall that $ \displaystyle {n\choose{m}} = \frac{n!}{m!(n-m)!} $
 
  • #3
Prove It said:
Maybe expanding the binomial coefficients would help. Recall that $ \displaystyle {n\choose{m}} = \frac{n!}{m!(n-m)!} $

Well I have tried that of course:
$$ \sum_{r=1}^k \frac{k!}{r!(k-r)!} \frac{(n-k-1)!}{(r-1)!(n-k-r)!} \stackrel{?}{=} \frac{(n-1)!}{(k-1)!(n-k)!} $$
But I don't know where to go from here since I still can't sum the left hand side. I also tried to prove it by induction but I failed to prove the induction step.
 
  • #4
VincentP said:
I'm having trouble proving the following identity (I don't even know if it's true):
$$\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\binom{n-1}{k-1}$$ $$\forall n,k \in \mathbb{N} : n>k$$
Thank you in advance for any help!
Vincent

Hi Vincent, :)

Thank you for submitting this problem, I enjoyed solving it. :)

Since, \(\binom{k}{r}=\binom{k}{k-r}\) we have,

\[\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\sum_{r=1}^{k}\binom{k}{k-r} \binom{n-k-1}{r-1}\]

Using the Pascal's rule we get,

\begin{eqnarray}

\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}&=&\sum_{r=1}^{k}\left[{k+1 \choose k-r+1}-{k \choose k-r+1}\right]\binom{n-k-1}{r-1}\\

&=&\sum_{r=1}^{k}{k+1 \choose k-r+1}\binom{n-k-1}{r-1}-\sum_{r=1}^{k}{k \choose k-r+1}\binom{n-k-1}{r-1}\\

\end{eqnarray}

Now use the Vandermonde's identity to get,

\[\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}={n \choose k}-\binom{n-1}{k}\]

Using the Pascal's rule again we get,

\[\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}={n-1 \choose k-1}\]

Kind Regards,
Sudharaka.
 
  • #5
For a completely different, combinatorial, way to look at this problem, suppose that you have $n-1$ objects, and you want to select $k-1$ of them. There are $n-1\choose k-1$ ways of making the selection.

Now suppose that $k$ of the objects are white, and the remaining $(n-1)-k$ are black. Then another way to select $k-1$ objects is as follows. First, choose a number $r$ between 1 and $k$ (inclusive). Then select $r-1$ black balls and $(k-1)-(r-1) = k-r$ white balls. The number of ways to do that is ${k\choose k-r}{(n-1)-k\choose r-1} = {k\choose r}{n-k-1\choose r-1}.$ Sum that from $r=1$ to $k$ to get the total number of ways to select $k-1$ objects.
 
  • #6
@ Sudharaka
Thank you very much for your explanation!
I have one question though: Doesn't Vandermonde's identity require the sum to start at r=0?

@Opalg
Thank you for your reply, that's a very interesting approach to the problem!
 
  • #7
VincentP said:
@ Sudharaka
Thank you very much for your explanation!
I have one question though: Doesn't Vandermonde's identity require the sum to start at r=0?

@Opalg
Thank you for your reply, that's a very interesting approach to the problem!

You are welcome. :) I have neglected an in between step that may have aroused the confusion.

\[\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\sum_{r=1}^{k}{k+1 \choose k-r+1}\binom{n-k-1}{r-1}-\sum_{r=1}^{k}{k \choose k-r+1}\binom{n-k-1}{r-1}\]

Substitute \(u=r-1\). Then the right hand side becomes,

\[\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\sum_{u=0}^{k}{k+1 \choose k-u}\binom{n-k-1}{u}-\sum_{u=0}^{k}{k \choose k-u}\binom{n-k-1}{u}\]

I hope this clarifies your doubts. :)

Kind Regards,
Sudharaka.
 
  • #8
Sudharaka said:
You are welcome. :) I have neglected an in between step that may have aroused the confusion.

\[\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\sum_{r=1}^{k}{k+1 \choose k-r+1}\binom{n-k-1}{r-1}-\sum_{r=1}^{k}{k \choose k-r+1}\binom{n-k-1}{r-1}\]

Substitute \(u=r-1\). Then the right hand side becomes,

\[\sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\sum_{u=0}^{k}{k+1 \choose k-u}\binom{n-k-1}{u}-\sum_{u=0}^{k}{k \choose k-u}\binom{n-k-1}{u}\]

I hope this clarifies your doubts. :)

Kind Regards,
Sudharaka.

Well not quite.
If you substitute the index of summation $ u=r-1 $ you have to change the lower as well as the upper bound of summation, because otherwise you change the number of summands. Therefore if you substitute $ u=r-1 $ we get:
$$ \sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\sum_{u=0}^{k-1}{k+1 \choose k-u}\binom{n-k-1}{u}-\sum_{u=0}^{k-1}{k \choose k-u}\binom{n-k-1}{u} $$ Which doesn't match Vandermonde's Identity anymore, because the upper bound of summation doesn't appear in the lower index of the binomial coefficant.

Kind Regards,
Vincent
 
  • #9
VincentP said:
Well not quite.
If you substitute the index of summation $ u=r-1 $ you have to change the lower as well as the upper bound of summation, because otherwise you change the number of summands. Therefore if you substitute $ u=r-1 $ we get:
$$ \sum_{r=1}^k \binom{k}{r} \binom{n-k-1}{r-1}=\sum_{u=0}^{k-1}{k+1 \choose k-u}\binom{n-k-1}{u}-\sum_{u=0}^{k-1}{k \choose k-u}\binom{n-k-1}{u} $$ Which doesn't match Vandermonde's Identity anymore, because the upper bound of summation doesn't appear in the lower index of the binomial coefficant.

Kind Regards,
Vincent

Note that,

\[\sum_{u=0}^{k-1}{k+1 \choose k-u}\binom{n-k-1}{u}-\sum_{u=0}^{k-1}{k \choose k-u}\binom{n-k-1}{u}=\sum_{u=0}^{k}{k+1 \choose k-u}\binom{n-k-1}{u}-\sum_{u=0}^{k}{k \choose k-u}\binom{n-k-1}{u}\]

since when \(u=k\) the right hand side is equal to zero. If you have any more questions about this please don't hesitate to ask. :)

Kind Regards,
Sudharaka.
 
  • #10
Sudharaka said:
Note that,

\[\sum_{u=0}^{k-1}{k+1 \choose k-u}\binom{n-k-1}{u}-\sum_{u=0}^{k-1}{k \choose k-u}\binom{n-k-1}{u}=\sum_{u=0}^{k}{k+1 \choose k-u}\binom{n-k-1}{u}-\sum_{u=0}^{k}{k \choose k-u}\binom{n-k-1}{u}\]

since when \(u=k\) the right hand side is equal to zero. If you have any more questions about this please don't hesitate to ask. :)

Kind Regards,
Sudharaka.

I think that clarifies everything, thanks so much.
Vincent
 
  • #11
VincentP said:
I think that clarifies everything, thanks so much.
Vincent

I am glad to be of help. :)

Kind Regards,
Sudharaka.
 

FAQ: Is This Binomial Coefficient Identity True?

What is the formula for calculating a sum with binomial coefficients?

The formula for calculating a sum with binomial coefficients is given by (x + y)^n = ∑k=0n (n choose k) xn-k yk, where n is the power of the binomial and k ranges from 0 to n.

How can I simplify a sum with binomial coefficients?

One way to simplify a sum with binomial coefficients is to use the Pascal's Triangle method. This involves writing the binomial coefficients in a triangular pattern, where each number is the sum of the two numbers above it. This method can help to quickly find the values of the coefficients and simplify the sum.

What is the relationship between Pascal's Triangle and binomial coefficients?

Pascal's Triangle is a visual representation of the coefficients in the expansion of a binomial. Each row in the triangle represents the coefficients in the expansion of the binomial raised to a certain power. The first and last numbers in each row are always 1, and the rest of the numbers are calculated by adding the two numbers above them.

Can binomial coefficients be negative?

No, binomial coefficients cannot be negative. They are always positive integers, and they represent the number of ways to choose k objects from a set of n objects, regardless of order.

In what real-life applications are binomial coefficients used?

Binomial coefficients have many applications in mathematics, statistics, and computer science. They are used in probability and combinatorics to calculate the number of possible outcomes in an experiment. They are also used in the binomial theorem to expand and simplify binomial expressions. In computer science, binomial coefficients are used in algorithms for solving various problems, such as finding the shortest path in a graph or determining the number of possible combinations in a set of data.

Similar threads

Replies
2
Views
1K
Replies
9
Views
1K
Replies
2
Views
2K
Replies
7
Views
2K
Replies
15
Views
2K
Replies
11
Views
1K
Replies
6
Views
1K
Back
Top