Re: xirr implementation
- From: Morten Welinder <mortenw gnome org>
- To: Jean Bréfort <jean brefort normalesup org>
- Cc: Schubert Tonarelli <schubert oktana io>, Gnumeric Mailing List <gnumeric-list gnome org>
- Subject: Re: xirr implementation
- Date: Sat, 12 Dec 2015 13:29:52 -0500
Gnumeric gets -80.59% too, at least for me. Jean: typo?
Note, that the input isn't particular reasonable: a loan of $15M being
repaid with three payments of $1.3M. That clearly requires a large
negative interest rate.
Xirr is based on finding roots of a high-degree polynomial. That is a
bit of an art form and there can be more than one solution. Finding
the roots is an iterative process -- no closed-form formula can exist
for degree of 5 or more.
In your 577 day example there could have been that many roots
and thus different answers. I mention this because in order to
get the result that the users wants, one really has to constrain
the roots to a range. Gnumeric uses a lower limit of -100% in
the search and an upper limit here of 57700%.
Maybe OO uses a different lower limit and thus cannot find the root.
Morten
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]