41 - Maximum Encyclopedia Overhang
Back to Topic List
Hi Kevin,
the test input data field is empty, although the checker has exact expected answers. Could you please populate it?
Thanks!
Thank you Kevin!
I have one more note. I use double type (which has 15-17 digit precision) for calculating the harmonic series (H(n)). For very large numbor of n I use the Euler-Mascheroni constant (gamma) for the approximation of H(n) lim(H(n)) -> ln(n) + gamma
For the expected answers my results differ in the last (8th) decimal digit. I tried truncation and rounding as well but there are always a few inputs which differ.
Maybe the issue is the same as we saw in other problems when the checker was not so permissive :) Could you please look into this?
Thank you in advance!
Hi Adam!
My first thoughts - even though 8 digits are being reported in the expected
answer, the actual checker for this problem does perform a comparison between submitted answers
and expected ones to verify that the differences between each pair is less than
1e-6, using those differences to determine pass/fail :)
However looking at your code I believe I see the potential issue - the
approximation you're referring to decreases its error as n increases, and as
it stands it looks like for the lowest value of n allowed to use that method in your
code will still yield error greater than 1e-6.
As an alternative to the straightforward fix (which should actually be
sufficient here), there are also options to add correction terms to that
approximation so that it is accurate to 1e-6 starting at much smaller values
of n... but that option shouldn't be strictly necessary here :)
Ok, I increased the lower bound where the approximation should take place. My solution is accepted now. Thanks.:)