Talk:Wieferich pair

Relation with Wieferich primes
I'm thinking of proposing a merger. Richard Pinch (talk) 10:44, 2 August 2008 (UTC)

Wieferich sequence
Start with a(1) any natural number (>1), a(n) = the smallest prime p such that (a(n-1))p-1 = 1 (mod p2) but p2 does not divide a(n-1)-1 or a(n-1)+1. It is a conjecture that every natural number a(1) makes this sequence become periodic, for example, let a(1) = 2:


 * 2, 1093, 5, 20771, 18043, 5, 20771, 18043, 5, ..., it gets a cycle: {5, 20771, 18043}.

Let a(1) = 83:


 * 83, 4871, 83, 4871, 83, 4871, 83, ..., it gets a cycle: {83, 4871}.

Let a(1) = 59 (a longer sequence):


 * 59, 2777, 133287067, 13, 863, 7, 5, 20771, 18043, 5, ..., it also gets 5.

However, there are many values of a(1) with unknown status, for example, let a(1) = 3:


 * 3, 11, 71, 47, ? (There are no known Wieferich primes in base 47).

Let a(1) = 14:


 * 14, 29, ? (There are no known Wieferich prime in base 29 except 2, but 22=4 divides 29-1 = 28)

Let a(1) = 39 (a longer sequence):
 * 39, 8039, 617, 101, 1050139, 29, ? (It also gets 29)

Do values for a(1) exist such that the resulting sequence does not eventually become periodic?

When a(n-1)=k, a(n) will be (start with k=2): 1093, 11, 1093, 20771, 66161, 5, 1093, 11, 487, 71, 2693, 863, 29, 29131, 1093, 46021, 5, 7, 281, ?, 13, 13, 25633, 20771, 71, 11, 19, ?, 7, 7, 5, 233, 46145917691, 1613, 66161, 77867, 17, 8039, 11, ...

— Preceding unsigned comment added by 101.15.67.1 (talk) 14:26, 30 July 2015 (UTC)

Rename article "Double Wieferich pair"
I suggest to rename this page to "Double Wieferich pair" or "Double Wieferich prime pair". Almost all sources dealing with these pairs call them "Double Wieferich pairs". --  Toshio   Yamaguchi  10:35, 27 August 2017 (UTC)

Wieferich sequence
Shouldn't this: "the smallest prime p such that a(n-1)p-1 = 1 (mod p)" read "... = 1 (mod p^2)"? Otherwise it's true for all p. --Bur (talk) 06:08, 25 June 2022 (UTC)