A question about factorials, nCr

  • Thread starter Thread starter davon806
  • Start date Start date
  • Tags Tags
    Factorials
AI Thread Summary
The discussion revolves around finding the smallest value of m such that C(2015, m) is an even number, concluding that m must be 32 based on the condition that (2016 - m)/m is even. The user expresses confusion about ensuring C(2015, m) remains an integer as m increases, particularly questioning scenarios where terms could lead to fractions. A key point clarified is that C(N, m) is always an integer for positive integers N and m, regardless of the size of m. The discussion also references Pascal's triangle and induction as methods to prove the integer nature of binomial coefficients. Overall, the mathematical principles confirm that C(2015, m) will not yield fractions for valid m values.
davon806
Messages
147
Reaction score
1

Homework Statement


Hi,I found this problem when I was reading a book,I knew the answer but there is one thing that I don't understand.Here is the question:

Given C(2015,m),find the smaller of m such that C(2015,m) is an even number

Homework Equations

The Attempt at a Solution


Here is the answer:(sorry for being quite messy)

Expand C(2015,m),which gives:

(2015)(2014)... (numerator)
m(m-1)(m-2)...(4)(3)(2)(1)...(2015-m)(2015-m+1)... (denominator)

After some arrangement it becomes:
(2015)(2014)...(2015-(m-1))
m(m-1)...(4)(3)(2)(1)

which is :
(2015)(2014)...(2016-m)
(1)(2)(3)(4)...(m)

In my book,the answer said the first even number occurs when (2016-m)/m is an even number
Therefore 2016/m - 1 must be even,then 2016/m must be odd.
Since 2016 = 2^5 * 63,the smallest number of m is 2^5 = 32

My question:
I know the answer is correct.However,let's just expand a few terms up to 2011 (i.e.m=5)

(2015)(2014)(2013)(2012)(2011)
(5) (4) (3) (2) (1)

which gives:

(403)(1007)(671)(503)(2011) = 2.75*(10^14)

Since m is small,the final product C(2015,m) has a high chance to be an integer as the factors in the denominator is quite small.
However,how can we ensure this is still the case when m is large(Just say m=32 given in my case),could it be a fraction instead of an even number?There might be terms like (1984/19) which makes the final term neither odd nor even.

Thanks very much.
 
Physics news on Phys.org
davon806 said:

Homework Statement


Hi,I found this problem when I was reading a book,I knew the answer but there is one thing that I don't understand.Here is the question:

Given C(2015,m),find the smaller of m such that C(2015,m) is an even number

Homework Equations

The Attempt at a Solution


Here is the answer:(sorry for being quite messy)

Expand C(2015,m),which gives:

(2015)(2014)... (numerator)
m(m-1)(m-2)...(4)(3)(2)(1)...(2015-m)(2015-m+1)... (denominator)

After some arrangement it becomes:
(2015)(2014)...(2015-(m-1))
m(m-1)...(4)(3)(2)(1)

which is :
(2015)(2014)...(2016-m)
(1)(2)(3)(4)...(m)

In my book,the answer said the first even number occurs when (2016-m)/m is an even number
Therefore 2016/m - 1 must be even,then 2016/m must be odd.
Since 2016 = 2^5 * 63,the smallest number of m is 2^5 = 32

My question:
I know the answer is correct.However,let's just expand a few terms up to 2011 (i.e.m=5)

(2015)(2014)(2013)(2012)(2011)
(5) (4) (3) (2) (1)

which gives:

(403)(1007)(671)(503)(2011) = 2.75*(10^14)

Since m is small,the final product C(2015,m) has a high chance to be an integer as the factors in the denominator is quite small.
However,how can we ensure this is still the case when m is large(Just say m=32 given in my case),could it be a fraction instead of an even number?There might be terms like (1984/19) which makes the final term neither odd nor even.

Thanks very much.

Fact of life: for positive integers N and m (with 0 ≤ m ≤ N), C(N,m) is always an integer.

For instance: C(N,0) = 1 is an integer; C(N,1) = N is an integer; C(N,2) = N(N-1)/2, and one of the factors in the numerator is even; C(N,3) = N(N-1)(N-2)/(3*2), and one of the factors in the numerator is even and one of them is a multiple of 3.

It becomes trickier for m larger than 3, but the result is still true.

Perhaps the easiest way is to do it is to first establish the recursion C(N,m) = C(N-1,m) + C(N-1,m-1), and then use induction on N. That is, given m, let N = m, m+1, m+2, ... .
 
Last edited:
There are a couple of ways to see it will always be an integer.
If you prove Pascal's relationship C(n,r)=C(n-1,r-1)+C(n-1,r) then inductively they are all integers.
Or consider that if a prime p occurs in some pattern in [1, r] then it occurs in a corresponding pattern in [n-r+1,n]. By that I mean that if p occurs to some power in the lower sequence then there is a corresponding place in the upper sequence where it occurs to at least the same power, and the correspondence is 1-1. That takes a bit of work to make rigorous.
 
Back
Top