Prove an identity with binomial coefficients

Click For Summary

Discussion Overview

The discussion revolves around proving the identity involving binomial coefficients: $\sum_{j=1}^{2n-1}\frac{(-1)^{j-1}j}{{2n \choose j }} = \frac{n}{n+1}$. Participants explore various approaches to tackle this problem, including proof by induction and numerical observations, while expressing the challenges they face.

Discussion Character

  • Exploratory
  • Technical explanation
  • Debate/contested
  • Mathematical reasoning

Main Points Raised

  • One participant suggests using proof by induction but finds it difficult to apply effectively.
  • Another participant references a related proof involving sums of reciprocals of binomial coefficients found on Mathematics Stack Exchange.
  • A participant shares numerical examples for small values of $n$, noting that the left side of the identity simplifies to the right side for $n=4$, but expresses that this does not simplify the problem overall.
  • There is an observation about pairing terms from each end of the sum, which results in a common denominator, but participants indicate that this insight does not lead to a resolution of the problem.
  • One participant acknowledges the complexity of the problem and expresses gratitude for the shared insights and numerical observations, indicating a sense of camaraderie in facing the challenge.

Areas of Agreement / Disagreement

Participants generally agree that the problem is difficult and that they have not found a straightforward approach to proving the identity. Multiple competing views and methods are presented, but no consensus is reached on a solution.

Contextual Notes

Participants mention the complexity of the notation in referenced papers and the challenges of applying certain mathematical techniques, indicating that there may be unresolved steps or assumptions in their reasoning.

lfdahl
Gold Member
MHB
Messages
747
Reaction score
0
Prove, that

$\sum_{j=1}^{2n-1}\frac{(-1)^{j-1}j}{{2n \choose j }} = \frac{n}{n+1}$

i have tried with proof by induction, but it is very difficult to use this technique.

I should be very glad to see any approach, that can crack this nut.
 
Physics news on Phys.org
lfdahl said:
Prove, that

$\sum_{j=1}^{2n-1}\frac{(-1)^{j-1}j}{{2n \choose j }} = \frac{n}{n+1}$

i have tried with proof by induction, but it is very difficult to use this technique.

I should be very glad to see any approach, that can crack this nut.
This looks like a difficult problem. The only help I could find online is a paper Alternating sums of the reciprocals of binomial coefficients. You might find that helpful if you can wade your way through the dense notation.

It is fascinating to see how the formula $\sum_{j=1}^{2n-1}\frac{(-1)^{j-1}j}{{2n \choose j }} = \frac{n}{n+1}$ works for small values of $n$. For example, when $n=4$ the left side is $\frac18 - \frac2{28} + \frac3{56} - \frac4{70} + \frac5{56} - \frac6{28} + \frac78$, which magically simplifies to $\frac45.$ Notice that if you pair together the terms from each end of the sum that have the same denominator, their sum has the constant numerator $2n$ (except that the middle term stands on its own, and its numerator is $n$). But that does not seem to make the problem any simpler.
 
sorry about my post, absolute nonsense mate
 
Fermat said:
sorry about my post, absolute nonsense mate
No problem at all! ;)

- - - Updated - - -

Opalg said:
This looks like a difficult problem. The only help I could find online is a paper Alternating sums of the reciprocals of binomial coefficients. You might find that helpful if you can wade your way through the dense notation.

It is fascinating to see how the formula $\sum_{j=1}^{2n-1}\frac{(-1)^{j-1}j}{{2n \choose j }} = \frac{n}{n+1}$ works for small values of $n$. For example, when $n=4$ the left side is $\frac18 - \frac2{28} + \frac3{56} - \frac4{70} + \frac5{56} - \frac6{28} + \frac78$, which magically simplifies to $\frac45.$ Notice that if you pair together the terms from each end of the sum that have the same denominator, their sum has the constant numerator $2n$ (except that the middle term stands on its own, and its numerator is $n$). But that does not seem to make the problem any simpler.

Thanks a lot, Opalg, for your thorough considerations and numerical observations. I did observe my self, that pairing the terms from each end of the sum results in a common denominator ($2n$), but I was not able to proceed and get any closer to a final result ...:(
Well, anyway, from your considerations its a "comfort" to know, that the problem in fact is difficult, and that I didn´t overlook any obvious approach. Thankyou once again!
 

Similar threads

  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 0 ·
Replies
0
Views
2K
Replies
1
Views
3K
  • · Replies 19 ·
Replies
19
Views
3K
  • · Replies 29 ·
Replies
29
Views
5K
  • · Replies 3 ·
Replies
3
Views
3K
Replies
2
Views
3K
  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 2 ·
Replies
2
Views
3K