It is straight forward to see that:

$$\tag0p_{\pi(n)+1}-p_{\pi(n+1)}=0 \operatorname{iff} \quad n \in {\{p_k-1:k \in \mathbb N}\}$$

however what is not as straightforward firstly, is to show that $n$ must be the lesser of a pair of twin primes if the same expression is equal to $2$:

$$\tag1p_{\pi(n)+1}-p_{\pi(n+1)}=2 \quad\operatorname{iff} \quad \exists k,j \in \mathbb N \quad\operatorname{s.t}\quad n=p_k=p_j-2$$

And furthermore, establish a proof for:

$$\tag2p_{\pi(n)+1}\equiv p_{\pi(n+1)}(\operatorname{mod}2)$$

Also an equality that I have been unable to determine the truth value of:

$$\Bigl\lfloor \frac{\ln(p_n)}{\ln(p_{\pi(n)})} \Bigr\rfloor=\Bigl\lfloor \frac{\lfloor\ln(p_n)\rfloor}{\ln(p_{\pi(n)})} \Bigr\rfloor$$

To simplify $(3)$ I defined a function $f$: $$f(n,k)=\Bigl\lfloor \frac{\ln(n+k)}{\ln(n)} \Bigr\rfloor-\Bigl\lfloor \frac{\lfloor\ln(n+k)\rfloor}{\ln(n)} \Bigr\rfloor$$

$$f(n_0,k-1)-f(n_0,k)=1$$ $$f(n_1,k-1)-f(n_1,k)=1$$ $$f(n_2,k-1)-f(n_2,k)=1$$ $$f(n,k-1)-f(n,k)=0 \quad\quad n_0 \lt n \lt n_1\land \quad\quad n_1 \lt n \lt n_2$$ Because $$n_2-n_1 \gt n_1-n_0$$ Are we able to assert that: $$n_2-n_1 \rightarrow \infty$$


$g(n,k)=\Bigl\lfloor \frac{\lfloor\ln(n+k)\rfloor}{\lfloor\ln(n)\rfloor} \Bigr\rfloor-\Bigl\lfloor \frac{\ln(n+k)}{\ln(n)} \Bigr\rfloor$

$g(n,k)=0 \operatorname{for} n \gt 7, k \gt 0$

Many thanks in advance.

  • $\begingroup$ Not sure I follow. Aside from $p=2$, all primes are odd. $\endgroup$
    – lulu
    Nov 7, 2018 at 21:13
  • $\begingroup$ Correct yes I meant to express $\tag1p_{\pi(n)+1}-p_{\pi(n+1)}$ appears to be always even $\endgroup$ Nov 7, 2018 at 21:15
  • $\begingroup$ I'm sorry I go very crazy when ever I think of Goldbach's so yes $(2)$ is very obvious I've just drunk too much tonight $\endgroup$ Nov 7, 2018 at 21:35

1 Answer 1


$(2)$ can only fail if one of the primes $p_{\pi(n)+1}$, $p_{\pi(n+1)}$ is even and the other is odd. As $\pi(n)\le \pi(n+1)\le \pi(n)+1$, this requires $\pi(n)=\pi(n+1)=1$. So both $n$ and $n+1$ must be $\ge 2$ and at the same time $\le 3$.

  • $\begingroup$ So it is as easy as simply stating the difference of two odd numbers is always even? $\endgroup$ Nov 7, 2018 at 21:22
  • $\begingroup$ I just don't understand if there exists a prime gap equal to any even number, and all sums of two primes are even, and there are infinite primes, how is Goldbach's Conjecture a conjecture? This is very strange for me $\endgroup$ Nov 7, 2018 at 21:28
  • $\begingroup$ ah I think I see now what is missing is the proof of the infinitude of twin primes $\endgroup$ Nov 7, 2018 at 21:32

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .