$ (\frac{n}{3})^{n}<n! < (\frac{n}{2})^{n}$

My appraoch so far

Idon't know for which inductionbase this formula would applyso I have startedwith the inductionstep:

$ (n+1)! = (n+1)n! > (n+1)(\frac{n}{3})^{n} = n^{n}(\frac{1}{3})^{n}(n+1)$

$(\frac{n+1}{3})^{n+1}= (\frac{n+1}{3})^{n}(\frac{n+1}{3})=(\frac{1}{3})^{n}(n+1)(\frac{(n+1)^n}{3})$

Only thing left to show would be $n^n \geq (\frac{(n+1)^n}{3})$

But I am not sure if I haven't done a mistake yet and hoq to prove this last inequality. Note that this excercise is from a chapter where biniommialcoefficients werenot introduced yet.

I also need to find the introductionbase and proof the other side of the inequality, is it possible to use the bernoulliinequality somewhere?


We need to know that $$2\le\left(\frac{n+1}{n}\right)^n<e<3$$ for all $n\ge1$.

So, if $n!<(n/3)^n$, then $$(n+1)!<(n+1)\frac{n^n}{3^n}<(n+1)\frac{n^n}{3^{n+1}}\left(\frac{n+1}{n}\right)^n=\left(\frac{n+1}3\right)^{n+1}.$$ The other inequality is similar.

  • $\begingroup$ I have got problems to proof the right inequalityof your prerequisite, I have already made a question about it. If you know how I can proofit using basic induction I wouldbe glad if you couldgive me an answer to it. My approach so far : $ b_n - b_{n+1} > 0 \rightarrow b_n > b_{n+1}\iff (1+\frac{1}{n})^{n+1}- (1+\frac{1}{1+n})^{n+2} > 0 \rightarrow (1+\frac{1}{n})^{n+1} > (1+\frac{1}{1+n})^{n+2}$ $\endgroup$ – RM777 Nov 7 '18 at 21:54
  • $\begingroup$ If the previous statementsare truethan by induction one can show that [so based on that $(1+\frac{1}{n})^n < 3 $ is true] $(1+\frac{1}{n+1})^{n+1}< (1+\frac{1}{n+1})^{n+2} < (1+\frac{1}{n})^{n+1}= (1+\frac{1}{n})(1+\frac{1}{n})^n < (1+\frac{1}{n})3 \rightarrow (1+\frac{1}{n+1})^{n+2} < (1+\frac{1}{n})3 \iff (1+\frac{1}{n+1})(1+\frac{1}{n+1})^{n+1} < (1+\frac{1}{n})3 \iff (1+\frac{1}{n+1})^{n+1} < \frac{1+\frac{1}{n}}{1+\frac{1}{n+1}}3$ $\endgroup$ – RM777 Nov 7 '18 at 21:57
  • $\begingroup$ Unfortunately the nominateris bigger than the denominater inthe last term so the estimate is not sufficient. $\endgroup$ – RM777 Nov 7 '18 at 22:00

Let us show that $a_n=\left(1+\frac{1}{n}\right)^n$ for $n\geq 1$ gives an increasing sequence bounded by $3$.

Increasing. The product of $k$ positive numbers can always be written as the $k$-th power of their geometric mean. In particular $$ 1\cdot\left(1+\frac{1}{n}\right)^n = \text{GM}\big(1,\underbrace{1+\tfrac{1}{n}}_{n\text{ times}}\big)^{n+1}\color{red}{<}\text{AM}\big(1,\underbrace{1+\tfrac{1}{n}}_{n\text{ times}}\big)^{n+1}=\left[\frac{1+n\left(1+\tfrac{1}{n}\right)}{n+1}\right]^{n+1} $$ by the AM-GM inequality. If we expand the RHS, we exactly get $a_n<a_{n+1}$.

Bounded by $3$. By the binomial theorem and the fact that $\binom{n}{k}\leq\frac{n^k}{k!}$ we have: $$ \left(1+\frac{1}{n}\right)^n = 1+\sum_{k=1}^{n}\binom{n}{k}\frac{1}{n^k}\leq 1+\sum_{k=1}^{n}\frac{1}{k!} $$ for any $n\geq 1$, hence $a_n$ is bounded by $1+\sum_{k\geq 1}\frac{1}{k!}$.
On the other hand, for any $k\geq 3$ we have $k!\geq 2\cdot 3^{k-2}$, hence $$ a_n \leq 1+1+\frac{1}{2}+\sum_{k\geq 3}\frac{1}{2\cdot 3^{k-2}}=\frac{11}{4}\color{red}{<3}. $$

Since any increasing and bounded sequence is convergent to its supremum, this shows that $$ \lim_{n\to +\infty}\left(1+\frac{1}{n}\right)^n $$ is a mathematical constant less than three, and with few efforts you may also prove that such mathematical constant is exactly $$ \sum_{k\geq 0}\frac{1}{k!} $$ a better-suited representation for numerical purposes, since such series is rapidly convergent. $\left(\sum_{k\geq 0}\frac{1}{2^k k!}\right)^2$ or $\left(\sum_{k\geq 0}\frac{(-1)^k}{k!}\right)^{-1}$ are even better.

Relations with the factorial. By defining $b_n$ as $\frac{n^n}{n!}$ we have $b_1=1$ and $$ \frac{b_{n+1}}{b_n} = \frac{(n+1)^{n+1}}{(n+1)!}\cdot\frac{n!}{n^n}=\left(1+\frac{1}{n}\right)^n, $$ hence $$ b_{n+1}=b_1\prod_{k=1}^{n}\frac{b_{k+1}}{b_k}=\prod_{k=1}^{n}\left(1+\frac{1}{k}\right)^k \in (2^n,3^n).$$ This gives a weak version of Stirling's inequality,

$$ \frac{(n+1)^{n}}{3^n}<n!\leq\frac{(n+1)^n}{2^n}.$$


It seems to me that this inequality is wrong.

This is the plot of the three functions enter image description here

And here a numerical example

$(\frac{5}{3})^5 = 12.86008$


$(\frac{5}{2})^5 = 97.65625$

It may be valid for some n or even for n > m but at least it is not valid for any n.

  • $\begingroup$ The original task was also to rather find theinduction base and not to do the induction. However once you have found an n for which the inequality holds you know that it will also hold for the successor. $\endgroup$ – RM777 Nov 7 '18 at 22:04
  • $\begingroup$ True, though the question is incomplete, isn't it? $\endgroup$ – Francesco Iovine Nov 8 '18 at 5:22

Since a sequence $a_n =(1+n^{-1})^n$ is increasing and its limit is equal $e<3$ we get $$(1+n^{-1})^n <3$$ for all $n\in \mathbb{N}$

  • $\begingroup$ Try taking logarithm and use stirling's approximation of n! You will get a strict inequality. Even you can start with $3^{-n }< e^{-n}< 2^{-n}$ $\endgroup$ – NewBornMATH Nov 6 '18 at 18:23
  • 1
    $\begingroup$ @NewBornMATH: How can you assume that Stirling's inequality is allowed when Euler's constant still has to be defined, reasonably? $\endgroup$ – Jack D'Aurizio Nov 6 '18 at 18:42
  • $\begingroup$ @Jack D' Aurizio sorry sir i did not get what you are saying i guess Stirling's inequality can be proved without euler's constant. math.stackexchange.com/questions/94722/stirlings-formula-proof $\endgroup$ – NewBornMATH Nov 7 '18 at 13:38
  • 1
    $\begingroup$ @NewBornMATH: if you write down Stirling's inequality/approximation, you'll see there is an $e=\lim_{n\to +\infty}\left(1+\frac{1}{n}\right)^n$ in it. $\endgroup$ – Jack D'Aurizio Nov 7 '18 at 13:48
  • $\begingroup$ @Jack D'Aurizio ohh ! Yes you are right i got it. I initially thought that you are talking about euler mascheroni constant.Apologies. $\endgroup$ – NewBornMATH Nov 7 '18 at 15:13

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.