PATCH: Advanced portfolio report revisions

[Mike or Penny Novack]

>We believe, but we are not sure, that the Microsoft-Money "Annual % Return" is the Internal Rate of Return (IRR).  See http://support.microsoft.com/kb/131664. We believe that the IRR is the "best" measure of return, because it allows any two investments to be meaningfully compared..  The calculation is non-trivial.  I recommend "Financial Calculation Programs for Linux" by James Shapiro at http://www.linuxjournal.com/article/2545.  He gives C, Java, and Perl programs for calculating IRR.
>  
>
Close. But just IRR by itself without including its standard deviation 
for the time period doesn't allow proper comparison of investments. Not 
that past measures of "risk" provide a certain measure for the future 
but it's the best that is available. The point here is that investment A 
with an IRR of M but a large deviation might not be a better than 
investment B with an IRR of N slightly smaller than M but with a much 
smaller deviation. The usual idea is that the real comparison gets made 
after adjusting for risk -- how much of the higher yielding would 
instead have to be invested at low yield very low risk to bring the 
blend to the same deviation -- and then use that blended yield.

The problem is that the deviation can't be simply calculated from the 
beginning and ending states but needs the history in between.

Michael