$$\prod_{n=3}^{\infty} \left(\;1-\frac{4}{n^2}\;\right)\;=\;\text{???}$$

I took the LCM and split the numerator as $(n+2)(n-2)$ and then took the product of the numerator and the denominator separately but I was not able to get the answer from that so can you please help me in what to do next.


Note that, by induction, we have $$F(K)=\prod _{n=3}^K \left(1-\frac{4}{n^2}\right) = \frac{(K+1) (K+2)}{6 K (K-1)}$$ for $K\geq3$.

It's them simple to see that $$\lim_{K\to\infty}F(K) = \frac{1}{6}.$$


A different solution. You can use the fact that $$\sin(\pi x)=\pi x \prod_1^\infty \left(1-\frac{x^2}{n^2} \right).$$

So your product is e qual to

$$\lim_{x \to 2} \frac{\sin(\pi x)}{\pi x (1-x^2)(1-x^2/4)}=1/6$$


$$ \prod_{n=3}^{\infty} \left( 1 - \frac{4}{n^2} \right) = \prod_{n=3}^{\infty} \left( \frac{(n-2)(n+2)}{n^2} \right) = \prod_{n=3}^{\infty} \left( \frac{n-2}{n} \right)\prod_{n=3}^{\infty} \left( \frac{n+2}{n} \right) = \lim_{n\rightarrow \infty} \frac{2^2}{4!} \frac{(n-2)!}{n!} \frac{(n+2)!}{n!} $$

So now we pull out Stirling's Approximation

$$ \frac{1}{3!} \lim_{n \rightarrow \infty} \frac{2 \pi \sqrt{n^2-4}\left( \frac{n-2}{e} \right)^{n-2}\left( \frac{n+2}{e} \right)^{n+2} }{2 \pi n \left( \frac{n}{e} \right)^{2n}} $$

Dividing out common terms we have:

$$ \frac{1}{3!} \lim_{n \rightarrow \infty} \frac{\sqrt{n^2-4}\left( \frac{n-2}{e} \right)^{n-2}\left( \frac{n+2}{e} \right)^{n+2} }{n\left( \frac{n}{e} \right)^{2n}} $$

We can take a natural logarithm and then exponentiate our answer to recover the original. So now consider:

$$ \lim_{n \rightarrow \infty} \frac{1}{2} \ln \left(\frac{n^2 - 4}{n} \right) + n \ln \left( \frac{(n+2)^2}{n^{n}(n-2)^2} \right)+n - \ln(3!) $$

Our problem then simplifies to:

$$ \lim_{n \rightarrow \infty} n \ln \left( \frac{(n+2)^2}{(n-2)^2} \right)+n - \ln(3!) - n \ln(n)$$

Which at limit yields:

$$ \lim_{n \rightarrow \infty} O(n) - O(n \ln n)$$

And so obviously tends to $-\infty$. This means that the original tends to 0.

So we conclude that the product tends to 0 (using our approximation).

An easier method.

Note that

$$ \frac{1 \times 2 ... \times n-2}{3 \times 4 ... \times n} = \frac{2}{(n-1)n} $$

On the other hand we have:

$$ \frac{5 \times 6 ... \times n+2}{3 \times 4 ... \times n} = \frac{(n+1)(n+2)}{3 \times 4} $$

So their combined product is

$$ \frac{2(n+1)(n+2)}{12(n-1)n} = \frac{1}{6} \frac{(n+1)(n+2)}{n(n-1)}$$

Now it's easy to prove $\lim_{n\rightarrow \infty} \frac{(n+1)(n+2)}{n(n-1)} = 1$ and you get the desired result.

  • 3
    $\begingroup$ But the answer is 1/6 $\endgroup$ – Gem Apr 20 '16 at 10:04
  • $\begingroup$ That means the approximation was too rough. $\endgroup$ – frogeyedpeas Apr 20 '16 at 10:05
  • $\begingroup$ Oh I see where it was too rough $\endgroup$ – frogeyedpeas Apr 20 '16 at 10:06
  • $\begingroup$ What is Stirling approximation $\endgroup$ – Gem Apr 20 '16 at 10:07
  • $\begingroup$ It's a way to approximate the factorial function. Sometime's it's useful when you want to get an asymptotic approximation of the factorial without being super accurate (such as in a limit). The problem is your multiplication meant that even the smaller parts that I was neglecting by the approximation affected the answer, and hence I made a mistake by using it. If you're curious you can read more about it here: en.wikipedia.org/wiki/Stirling%27s_approximation $\endgroup$ – frogeyedpeas Apr 20 '16 at 10:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.