I'm trying to show that $\exists \,\varepsilon >0\mid\forall n>N\in\mathbb{N}$ such that: $$\left|\frac{6n^3+5n-1}{2n^3+2n+8}-3\right| < \varepsilon$$ Let's take $\varepsilon = 1/2$:

$$\left|\frac{6n^3+5n-1}{2n^3+2n+8}-3\right| = \left|\frac{6n^3+5n-1 - 6n^3 -6n -24}{2n^3+2n+8}\right| < \left|\frac{-n -25}{2n^3+2n+8}\right| = \left|-\left(\frac{n +25}{2n^3+2n+8}\right)\right|$$

Since $|-x| = |x|$:

$$\left|-\left(\frac{n +25}{2n^3+2n+8}\right)\right| = \left|\frac{n +25}{2n^3+2n+8}\right|<\varepsilon $$

This is the part which I have trouble with. Here, I always end up with my minimum-required index to be smaller than zero, which seems peculiar to me:

$$\left|\frac{n +25}{2n^3+2n+8}\right| < |n+25| = n+ 25 < \varepsilon = 1/2$$ From here, I will get that my $N$ will be less than zero, which means that all the elements of the sequence are inside of my given epsilon environment, but I know that's not true since $a_1 = 5/6 < 3 - \varepsilon $, so what I did do wrong? Is it because I completely removed the denominator? If so, why does that break the inequality?

  • 3
    $\begingroup$ Yes $\left|\frac{n+25}{2n^3+2n+8}\right|\lt|n+25|$ but this will not allow you to solve the limit in question. I suggest writing instead $\left|\frac{n+25}{2n^3+2n+8}\right|\lt|\frac{1}{n^2}|$ for large enough $n$. $\endgroup$ Apr 16, 2019 at 21:54
  • 2
    $\begingroup$ You don't want to show that there exists $\varepsilon$, you want to show that any positive $\varepsilon$ satisfies the property. $\endgroup$
    – Git Gud
    Apr 16, 2019 at 21:58
  • $\begingroup$ See this for a similar example, and also for general strategy. $\endgroup$
    – Git Gud
    Apr 16, 2019 at 21:59
  • $\begingroup$ It seems that you want $\varepsilon$-$\delta$ (or $\varepsilon$-$n_0$) proof - which suggest that the (epsilon-delta) tag might be suitable here. $\endgroup$ Jun 5, 2019 at 14:19

4 Answers 4


Don't worry about $\epsilon$. Instead, try to get the difference in a simple form by assuming $n$ is as large as you need. Then getting $n$ is much simpler.

Using your calculations:

$\begin{array}\\ \left|\dfrac{6n^3+5n-1}{2n^3+2n+8}-3\right| &= \left|\dfrac{6n^3+5n-1 - 6n^3 -6n -24}{2n^3+2n+8}\right|\\ &= \left|\dfrac{-n -25}{2n^3+2n+8}\right|\\ &= \left|\dfrac{n +25}{2n^3+2n+8}\right|\\ &\le \left|\dfrac{2n}{2n^3}\right| \qquad\text{if } n \ge 25\\ &= \left|\dfrac{1}{n^2}\right|\\ &\lt \epsilon \qquad\text{if } n \gt \frac1{\sqrt{\epsilon}}\\ \end{array} $

Therefore the difference is within $\epsilon$ if $n \gt \max(25, \frac1{\sqrt{\epsilon}})$.

You don't have to get the best possible $n$ - just showing one exists is good enough.

  • $\begingroup$ Well, I worry a little bit about $\varepsilon$, when the first sentence of the OP starts with "$\exists \,\varepsilon >0\mid\forall n>N\in\mathbb{N}$". $\endgroup$ Apr 16, 2019 at 22:12
  • $\begingroup$ $N = 1/\sqrt{\epsilon}$. $\endgroup$ Apr 17, 2019 at 4:16

Note that for $n\geq 1$ we have $n+25\leq n+25n=26n$ and $2n^3+2n+8\geq 2n^3$ so you have that $$\left|\frac{n+25}{2n^3+2n+8}\right|\leq \frac{26n}{2n^3}=\frac{13}{n^2}$$

Can you take it from here?


I do it like this:

$\forall n \ge 1, \; \dfrac{6n^3 + 5n - 1}{2n^3 + 2n + 8} = \dfrac{6 + 5n^{-2} - n^{-3}}{2 + 2n^{-2} + 8n^{-3}}; \tag 1$

since it is easy to see that

$\displaystyle \lim_{n \to \infty} (6 + 5n^{-2} - n^{-3}) = 6, \tag 2$


$\displaystyle \lim_{n \to \infty} (2 + 2n^{-2} + 8n^{-3} ) = 2, \tag 3$

we have

$\displaystyle \lim_{n \to \infty} \dfrac{6n^3 + 5n - 1}{2n^3 + 2n + 8} = \lim_{n \to \infty} \dfrac{6 + 5n^{-2} - n^{-3}}{2 + 2n^{-2} + 8n^{-3}} = \dfrac{\lim_{n \to \infty}( 6 + 5n^{-2} - n^{-3})}{\lim_{n \to \infty} ( 2 + 2n^{-2} + 8n^{-3}) } = \dfrac{6}{2} = 3. \tag 4$

Of course, such an answer, though perfectly rigorous and on point, is bound to be less than satisfactory to readers who want, as the text of the question indicates, to see the $\epsilon$-$N$ mechanism operate in detail; what I have done here is simply invoke standard and elementary results on the behavior of limits; we may also cast things into $\epsilon$-$N$ form if we write

$\dfrac{6 + 5n^{-2} - n^{-3}}{2 + 2n^{-2} + 8n^{-3}} - 3$ $= \dfrac{ 6 + 5n^{-2} - n^{-3} - 3( 2 + 2n^{-2} + 8n^{-3})}{2 + 2n^{-2} + 8n^{-3}}$ $=\dfrac{5n^{-2} - n^{-3} - 6n^{-2} - 24n^{-3}}{2 + 2n^{-2} + 8n^{-3}} = \dfrac{-n^{-2} - 25n^{-3}}{2 + 2n^{-2} + 8n^{-3}} ; \tag 5$

$\left \vert \dfrac{6 + 5n^{-2} - n^{-3}}{2 + 2n^{-2} + 8n^{-3}} - 3 \right \vert = \left \vert \dfrac{-n^{-2} - 25n^{-3}}{2 + 2n^{-2} + 8n^{-3}} \right \vert; \tag 6$

it is evident via inspection of the right-hand side of this equation that, given any $\epsilon > 0$ there exists a sufficiently large $N \in \Bbb N$ that

$n > N \Longrightarrow \left \vert \dfrac{-n^{-2} - 25n^{-3}}{2 + 2n^{-2} + 8n^{-3}} \right \vert < \epsilon, \tag 7$

which in light of (6) by definition implies

$\displaystyle \lim_{n \to \infty} \dfrac{6 + 5n^{-2} - n^{-3}}{2 + 2n^{-2} + 8n^{-3}} = 3, \tag 8$

completing an $\epsilon$-$N$ demonstration of the requisite limit.

It appears to me that our OP Fatima Ens' principal error lies in confusing the roles of $\exists$ and $\forall$ in the definition of limit; the quantifying symbol string should read

$\forall \epsilon \exists N \mid \forall n > N \; \text{and so forth}; \tag 9$

if this correction is accepted and the remaining statements bought into accord with this (corrected) version, the chances of completing a successful proof are greatly enhanced.


I don't understand why everyone is making this so long and drawn out.

You only need remember this: as $n\to \infty$; $t\geq0\implies n^{t+1}-n^t\to\infty$

In other words, given $\lim_{n\to\infty}(\frac{P(n)}{Q(n)})$ where $P$ and $Q$ are polynomials, only the largest powers (as long as they are positive) and their coefficients count.

So: $$\lim_{n\to\infty}\bigg[\frac{6n^3+5n-1}{2n^3+2n=8}\bigg]=\lim_{n\to\infty}\bigg[\frac{6n^3}{2n^3}\bigg]=\lim_{n\to\infty}[3]=3$$


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.