Dear Neil,
Thank you for writing. I will consider Seqfan subscription!
I would like to clarify the current state of comments in OEIS regarding sequences A068499 and A027423, and explain the complete mathematical result I have now proved.
CURRENT STATE OF MY OEIS COMMENTS:
For A068499 (numbers m such that m! is not divisible by m+1): The comment states "Also n such that tau((n+1)!) = 2* tau(n!)" (which is not mine I think). This phrasing is ambiguous: it reads like a complete characterization ("all n such that..."), but it could mean only one direction (if n is in this sequence, then tau((n+1)!) = 2tau(n!)), which would have been clearer if written as "Also, if n is in this sequence then tau((n+1)!) = 2tau(n!)".
For A027423 (number of divisors of n!): My 2002 comment states "It appears that a(n+1) = 2*a(n) if n is in A068499." This explicitly conjectures only one direction (sufficiency).
MY COMPLETE RESULT (2025):
I have now proved the full equivalence (if and only if): tau((n+1)!) = 2tau(n!) if and only if n is in A068499, equivalently: tau((n+1)!) = 2tau(n!) if and only if n = 3 or n+1 is prime.
THE TWO DIRECTIONS:
- Sufficiency: If n is in A068499, then tau((n+1)!) = 2*tau(n!). This direction is easy: Wilson's theorem tells us that n in A068499 means n+1 is prime or n=3, and the result follows from multiplicativity of tau (for primes) or direct computation (for n=3).
- Necessity: If tau((n+1)!) = 2*tau(n!), then n is in A068499. This is the nontrivial direction.
THE NONTRIVIAL PART (necessity):
Wilson's theorem gives a dichotomy: either (n+1) does not divide n! (Case 1: n+1 prime or n=3, which gives the desired conclusion), or (n+1) divides n! (Case 2: n+1 composite and n+1 not equal to 4, which must be ruled out).
In Case 2, one must prove that the equation tau((n+1)!) = 2*tau(n!) cannot hold. The proof uses Legendre's formula for prime valuations in factorials and an asymptotic argument (splitting by largest prime factor) showing the product is strictly less than 2 for all composite n+1 not equal to 4.
This completes my 2002 conjecture into a full theorem (if and only if) and clarifies the ambiguous comment on A068499.
I have written a complete proof in a short note that I can provide.
Best regards,
Benoit