"Stopping" A System
by John Sweeney
Once you get hold of an idea, it's often tough to get rid of it, especially if it shows no promise
whatever. It tugs at the back of the mind—what was it about the original insight? What was valid? What
was bogus? I hate to admit that my subconscious could have produced a bummer. Closing last month's
Settlement, I mused, "Reasonable progress, but what next? Could something simple work profitably?"
My first thought on returning to the subject and reviewing Figures 1a and 1b was that the losses were
pretty big. Those needed to be reduced in size, if not number. Second, November and December had been
removed from the graph! Looking at the work sheet, the data did indeed end in October now instead of
December. Loading the same data file into a fresh model showed the same result. I checked the other data
files, but they were intact. I put in a puzzled call to developer Jim Payne and decided to move on with the
remaining data, since it covered all the trades the original trading model had taken.
In the profit and loss column, I noted several losses over $1,000. I like to keep losses to $600 to $700, so
these were obvious candidates for stops. Looking down the max loss column, I noted only one winner
with a loss of more than $1,000 —a trade that eventually produced a $6,006 profit. Only one losing trade
would have been worse off by cutting its loss at $1,000. I scribed a square about its $1,150 max loss.
Parenthetically, System Writer Plus's (SWP) manual is mute on whether the max loss is an intraday figure
or a figure based on the closes, and thus another curious call to the support folks. Busy, busy, busy.