I post the following observation for information only:
recently Bernie and Tom posted their returns for
UOPIX YTD. that of course got me looking to see how
my simulation would do.
as reported, my simulated results returned 18% ytd on
UOPIX, on aa AIM sheet which started on 1/3/00.
however, today I got looking at my sheet which simulates
UOPIX back to 3/98 and found that in the same period
(1/3/00 - ytd), I acheived a 61% return.
so, 61% on a mature sheet and 18% on a new sheet. I likied the mature sheet better, so what I did was this:
I initialzed my new sheet on 1/3/00 with the same
cash reserves as existed on my mature sheet (which was about 39% cash at that time). I then calculated the % moves for each week on the existing (mature) sheet, and applied those % moves to the new sheet (rather than follow the signals which came directly from the new sheet). the %move was calculated (from the riedeman sheet) as the "Action Taken Value" / "Stock Value". "action taken value" is the amount actually exchanged on the trade.
the result: the new sheet (following the signals from the
old sheet) now returned 59% ytd...
I then tested this idea with various start dates, and even eliminated the last 8 weeks or so when I thought my existing
sheet had traded exceptionally well. no matter what I
did, my existing, mature sheet beat newly started sheets.
I then compared my ytd results between my two year UOPIX
sheet and my 11 year NDX sheet [in which I attempted to
simulate the UOPIX by doubling all moves], and in this
case I found the ytd returns to be comprable; meaning
that once a sheet reaches its peak performance, it doesnt get any better with the passge of time - it will, however,
beat the pants off the new sheet - which means that over time the new sheet will produce much greater returns as they
are compounded
the implications here are quite outstanding for starting
a new AIM security. If anyone out there would want to test
some other security in a similar manner I would be most interested in the results
Bob Wallace |
