7
$\begingroup$

Suppose $p$ is a prime such that $p + 2$ is also prime, and nothing else is known about $p$.

Is there any reason to think that this affects the probability that $p$ is also a Sophie Germain prime? Other than the fact that $p \pmod 6$ would have to be $5$? It would seem that the answer has to be no, but I have some perplexing computational results that I would like to account for.

$\endgroup$
4
  • 3
    $\begingroup$ What exactly is your numeric evidence? $\endgroup$ Mar 31, 2014 at 7:01
  • 5
    $\begingroup$ It's not just mod 6; there is a bias modulo every small prime. For instance, mod 5, a prime p is equal to 1,2,3,4 mod 5, and needs to avoid 2 mod 5 to have any chance to be a Sophie Germain prime (unless it actually is 2). But if p is a twin prime, it avoids 3 mod 5 (if it isn't actually 3), which slightly raises the chance that it is equal to 2 mod 5 instead. $\endgroup$
    – Terry Tao
    Mar 31, 2014 at 20:01
  • 6
    $\begingroup$ The asymptotics from the Hardy-Littlewood prime tuple conjecture can be used to predict the number of tuples of the form p, p+2, 2p+1 that are simultaneously prime, and this can lead to a prediction of the precise bias in the probability that 2p+1 is prime if one conditions on the event that p+2 is prime also. $\endgroup$
    – Terry Tao
    Mar 31, 2014 at 20:03
  • $\begingroup$ See COMMENTS by John W. Nicholson, on Dec 14 2013 at oeis.org/A168421 . $\endgroup$ Apr 10, 2014 at 2:45

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.