login

Year-end appeal: Please make a donation to the OEIS Foundation to support ongoing development and maintenance of the OEIS. We are now in our 61st year, we have over 378,000 sequences, and we’ve reached 11,000 citations (which often say “discovered thanks to the OEIS”).

A121719
Strings of digits which are composite regardless of the base in which they are interpreted. Exclude bases in which numbers are not interpretable.
2
4, 6, 8, 9, 20, 22, 24, 26, 28, 30, 33, 36, 39, 40, 42, 44, 46, 48, 50, 55, 60, 62, 63, 64, 66, 68, 69, 70, 77, 80, 82, 84, 86, 88, 90, 93, 96, 99, 100, 110, 112, 114, 116, 118, 120, 121, 130, 132, 134, 136, 138, 140, 143, 144
OFFSET
1,1
COMMENTS
"Think of these as polynomials. E.g. 121 is the polynomial n^2+2n+1. There are three cases:
"(1) If the coefficients (digits) all have a common factor, the result will be divisible by that factor.
"(2) If the polynomial can be factored, the numbers will be composite. n^2+2n+1 = (n+1)^2, so it is always composite.
"(3) Otherwise, look at the polynomial modulo primes up to its degree. For example, 112 (n^2+n+2, degree 2) modulo 2 is always 0, so it is always divisible by 2.
"Note that condition (1) is really a special case of condition (2), where one of the factors is a constant.
"If none of the above conditions apply, the polynomial will (probably) have prime values."
From Iain Fox, Sep 02 2020: (Start)
lim_{k->infinity} (1/k)*Sum_{i=1..k} a_c(i) > .3 if it exists, where a_c(n) is the characteristic function of a(n) (1 if n is in a(n), otherwise 0).
If the Bunyakovsky conjecture is true, the list of reasons a number is in this sequence detailed by Franklin T. Adams-Watters above is a complete list.
If the Bunyakovsky conjecture and the Extended Riemann Hypothesis are true, the above limit equals 4340435807/13235512500 = 0.3279386... (proof by Ravi Fernando in link by Iain Fox).
All members of A008592 except 1 and 10 are in this sequence.
(End)
EXAMPLE
String 55 in every base in which it is interpretable is divisible by 5. String 1001 in base a is divisible by a+1. Hence 55 and 1001 both belong to this sequence.
PROG
(PARI) is(n)=if(n<10, return(!isprime(n)&&n>1)); if(content(n=digits(n))>1, return(1)); if(vecsum(factor(n*=vectorv(#n, i, x^(#n-i)))[, 2])>1, return(1)); forprime(p=2, #n-1, for(x=1, p, if(eval(n)%p, next(2))); return(1)); for(x=vecmax(Vec(n))+1, +oo, if(isprime(eval(n)), return(0))) \\ Iain Fox, Aug 31 2020
CROSSREFS
Supersequence: A002808.
Sequence in context: A075243 A024370 A367176 * A267509 A162738 A366826
KEYWORD
nonn,easy,base
AUTHOR
Tanya Khovanova, Sep 08 2006
EXTENSIONS
More terms from Franklin T. Adams-Watters, Sep 12 2006
STATUS
approved