[seqfan] Re: Colorless integers

[Robert Gerbicz]

2011/2/22 N. J. A. Sloane <njas at research.att.com>

> Start at n, let a(n) be the smallest number in the cycle if it
> goes into a cycle, or -1 if it diverges. Is this sequence
> in the OEIS? and if not, you know what to do!
> Neil
>
> _______________________________________________
>
> Seqfan Mailing list - http://list.seqfan.eu/
>

My conjecture is that for every n in O(log(n)^2) iteration the sequence goes
to a cycle. The idea behind this that by O(1/log(n)) probability F(F(n))=n
is true.
Computations up to n=10^5 supports my conjecture.