[seqfan] Re: Reciprocal Recaman

[M. F. Hasler]

Oops! They differ from a(20) on.... the sign went wrong, should obviously be
s-=(-1)^(n*s<1)/n

Apologies... Punishment: I'll add the remarks to A231693...
Maximilian

> One could also xref
> A203811: Denominators of s(i) = s(i-1) - (1/i)*sign(s(i-1)) with s(1) = 1.
> based on a very similar idea (but which differs (more) often).
> Maximilian
> PS: it seems the two coincide at least up to the 43000+ digit a( n=10^5 ):
> {s=t=0;for(n=1,10^5,denominator(s+=(-1)^(n*s<1)/n)!=denominator(t+=1/n)&&return(n))}
> does not return...
>
> On Fri, Nov 15, 2013 at 10:06 PM, Neil Sloane <njasloane at gmail.com> wrote:
>> Oops! The denominators are A231693...

>>>> On Fri, Nov 15, 2013 at 4:51 AM, Don Reble <djr at nk.ca> wrote:
>>>>
>>>> > f(0) = 0
>>>> >> f(n) = f(n-1) - 1/n if >= 0, else f(n-1) + 1/n.
>>>> >>
>>>> >> For which n do we have f(n-2) > f(n-1) > f(n)?
>>>> >>