On Fri, Nov 13, 2015 at 10:12:45PM +0000, Wm... wrote: > I installed 2.2.7 on a spare Win XP system and a test suggests the IRR > figures are reasonable.
I am on linux only systems and installing the needed (old) libraries for 2.2.7 looks painful if even possible. I will try again tomorrow. > > I used the flow in > === > https://support.office.com/en-us/article/IRR-function-64925eaa-9988-495b- > b290-3ad0c163c1bc > === > rather than real data as my example and get > > an IRR of 8.66% vs their 8.7% after 5 years > an IRR of -2.12% vs their -2.1% after 4 years > an IRR of -44.30% vs their -44.4% after 2 years (no user guesses) > > which is close enough not to matter. I'd lean towards the code being correct as well. There are iterative techniques in play here and choosing a stopping criterion is needed; there may be a difference in how the code acts vs. Excel, and even in what Excel says it is doing and what it is doing in actuality. > > Generally I'm now thinking this might be a worthwhile addition. > > A thought: presuming the IRR is a good number should it be added to the > extant Advanced Portfolio report or should we make a new report? I'm > wondering if the Advanced Portfolio report isn't getting overloaded, people > are moaning about corner cases already, I think. I have been thinking about this more and I think one way to avoid corner cases with IRR is to simply make it calculate XIRR. This handles the non-dividend contribution case easily. With IRR you must have regular contributions/withdrawals which makes things less precise. The calculations are not very different, as you can imagine. As for putting it in a different report, perhaps, To my mind it may depend on how much configuration IRR/XIRR needs at the report-level. Regards, Thomas
signature.asc
Description: PGP signature
_______________________________________________ gnucash-devel mailing list [email protected] https://lists.gnucash.org/mailman/listinfo/gnucash-devel
