QUOTE(dasecret @ Mar 25 2016, 02:14 PM)
You know what, when I first read his example I have the same thought as you.... but, being a nerd bean counter as I am, put in spreadsheet and calculate lor
[attachmentid=6237477]
And this is what I got... the IRR is slightly positive, and hence my comment that it doesn't weight average returns but takes the time value of return instead
There is such a thing as GIGO - garbage in, garbage out. [attachmentid=6237477]
And this is what I got... the IRR is slightly positive, and hence my comment that it doesn't weight average returns but takes the time value of return instead
The extreme numbers broke the function! Try changing the 'time value' from 1/1/2015 to any date, and the XIRR remains the same...
Here's the logic on what the numbers are saying:
$10 gains on $10 invested is ROI 100%, and since the time period is 1 year, the IRR (or CAGR) is also 100%.
$1000 lost on $20,000 is -5% ROI, and since the time period is 1 day, the CAGR is -100%.
The projected lost (in one year) is very close to 100%, since if we lose 5% a day, we will lose NEARLY all in several weeks. We can't lose more than $20,000 (or in other words, less than -100%), because the maths is not a simple 'minus' ie. 20k - 1k - 1k - 1k ... but 20k x 95% x 95% x 95%....
Even if we lost 10% a day, which is $2000, the projected lost will still be capped at -100%.
Putting the 2 investments of 'extreme' figures and dates together into the XIRR function, it resulted in a confusing/misleading IRR of 0%.
But using cow sense and being aware of GIGO, any reasonable investor would ignore the initial $10, and think and calculate solely on the $20,000...
Mar 25 2016, 03:46 PM

Quote
0.0908sec
0.58
7 queries
GZIP Disabled