- --- In primenumbers@yahoogroups.com, "mikeoakes2" <mikeoakes2@...> wrote:

> Ok, let's try and spell it out.

"The prime sequence has no "memory""

>

> Start with an arbitrary (and large, so PNT applies) integer N.

>

> On either side of N, an integer has a chance 1/log(N) of being prime;

> so if one travels (in either direction), you have to go on average log(N) before finding a prime.

>

> p1 < N <=p < p2.

>

> From N, travel downwards: it takes log(N) to find p1, so av. (N-p1)=log(N).

> From N, travel upwards; it takes log(N) to find p, so av. (p-N)=log(N).

> The prime sequence has no "memory", so, continuing upwards, from p to p2 takes another log(N), so av. (p2-p)=log(N).

> Adding all 3 averages gives C=3.

>

> Mike

>

If this were true, then we would see the Ramanujan primes ~ 3n instead of ~ 2n. http://arxiv.org/abs/1105.2249

The "memory" is in a way a type of "feedback" by the prior primes to the future gap, and is related to the number of Ramanujan primes, n. See the corollary at http://en.wikipedia.org/wiki/Ramanujan_prime and think about how these log N's add up between x/2 and x. Note also Generalized Ramanujan primes and how they relate. Also note, how sharp the range is between [p_k,2*p_(i-n)]. This is where p_i is found. - --- In primenumbers@yahoogroups.com, "mikeoakes2" <mikeoakes2@...> wrote:

> Ok, let's try and spell it out.

"The prime sequence has no "memory""

>

> Start with an arbitrary (and large, so PNT applies) integer N.

>

> On either side of N, an integer has a chance 1/log(N) of being prime;

> so if one travels (in either direction), you have to go on average log(N) before finding a prime.

>

> p1 < N <=p < p2.

>

> From N, travel downwards: it takes log(N) to find p1, so av. (N-p1)=log(N).

> From N, travel upwards; it takes log(N) to find p, so av. (p-N)=log(N).

> The prime sequence has no "memory", so, continuing upwards, from p to p2 takes another log(N), so av. (p2-p)=log(N).

> Adding all 3 averages gives C=3.

>

> Mike

>

If this were true, then we would see the Ramanujan primes ~ 3n instead of ~ 2n. http://arxiv.org/abs/1105.2249

The "memory" is in a way a type of "feedback" by the prior primes to the future gap, and is related to the number of Ramanujan primes, n. See the corollary at http://en.wikipedia.org/wiki/Ramanujan_prime and think about how these log N's add up between x/2 and x. Note also Generalized Ramanujan primes and how they relate. Also note, how sharp the range is between [p_k,2*p_(i-n)]. This is where p_i is found.