cp's OEIS Frontend

Follow same procedure that is used to produce the lucky numbers A000959 except use primes instead of natural numbers.

Start with natural numbers, apply sieve of Eratosthenes, then sieve of Ulam. This is an example of composition of sieve operators. Circa 1955, Polish mathematician Stanislaw Ulam (1909-1984) identified a particular sequence which he designated "lucky numbers," which share many properties with primes (density, equivalent of twin primes, equivalent of Goldbach's conjecture). Other "random primes" which generalize the lucky numbers not only almost always satisfy the prime number theorem but also the Riemann Hypothesis. What can be said about composition of such "random primes"? - Jonathan Vos Post, May 08 2007

There is a slight ambiguity, arising from the first step of Ulam's sieve, which is to delete every second number, while in the remainder of the procedure, one deletes every v(k)-th term from the current vector v, with k=2,3,4... (but not k=1 in the 1st step). The present sequence is obtained by deleting in the first step every 2nd prime (thus using k=1 in the first step). - M. F. Hasler, Sep 23 2013

Originally arose as "Lucky numbers generated from squarefree numbers" under the hypothesis that Ulam's sieve (the one used to produce lucky numbers) ignores the values of the terms.

Result of application of "Ulam's sieve" (used to produce the lucky numbers A000959) to the noncomposite numbers A008578 instead of the naturals A000027.

The "Lucky numbers generated from primes", cf. A059987 = (2, 5, 11, 17, ...), form a disjoint subset, because they are a subset of the odd-indexed primes, while the present sequence is the union of {1} and a subset of the even-indexed primes. - M. F. Hasler, Sep 24 2013