Thinking on Oppermann and Goldbach conjectures, it came to me the following conjecture that mixes some aspects of both of them:
Let $n\notin\mathbb{P}$; then, there exist some prime numbers $p$ and $q$ such that $n^2-n<p<n^2<q<n^2+n$ such that $p+q=2n^2$
I have been checking first cases "by hand", and I would like to know if there exists an "easy" way to check if the conjecture is true or has some inmediate counterexample. I tried programming something in Excel but it was too messy.
Thanks in advance!
My gut feeling was there was going to be counterexamples. Essentially we want to know if for every $n$ there is an $i\in (0,n]$ such that $n^2-i$ and $n^2+i$ are both primes.
My very sloppy heuristic is that if we assume that a number in the range $(n^2-n,n^2+n)$ is prime with probability $1/\log(n)$ then the probability that both $n-i$ and $n+i$ are prime is $1/\log(n)^2$, so the chance it doesnt happen for any of our pairs is $(1-\frac{1}{\log(n)}^2)^{n}$ which seems to converge to $0$ quickly.
In any case I tested it for $n$ up to $10,000$ and these are the values of $n$ for which I got no solutions, we should probably use a sieve if we want to check for higher $n$, it seems for higher values the chance of none of the pairs working is indeed very small.
no solutions for $n = 5, 7, 11, 17, 20, 22, 23, 24, 31, 34, 37, 44, 49, 50, 57, 58, 62, 67, 73, 77, 82, 107, 118, 128, 131, 133, 142, 211, 277, 290, 334, 419$.
In particular $20$ seems to be the smallest non-prime counterexample.
c++ code: