Figure 6.17 The average profit over each monthly interval was substantially better without the trailing stop, suggesting that many trades were cut off prematurely. The unmarked bars are for 2 months, 4 months, and 6 months.
200 Equity Curve Analysis
the smoothness of the equity curve. For example, the standard error with the trailing stop was 3 percent higher, at $4,087, even though profits plummeted nearly 70 percent.
Deleting the trailing stop and adding an exit on the twentieth day of entry increased the reported profit to $14,950 (versus $7,500 with the trailing stop) with a profit factor of 1.27, and drawdown of-$11,325. These data are virtually the same as those with a $5,000 initial stop. Hence, there should be little change in portfolio level performance as a result of adding this exit.
The new SE was 10 percent smaller, at $3,781, but the performance over the intervals was comparable to tests without the stop. Hence, adding this exit produced little improvement, but cost a 40-percent drop from $24,000 in potential profits. Although not shown here, the proportion of profitable intervals dropped about 10 percent, another strike against this exit strategy.
A number of other exit strategies yielded similar results: none improved smoothness of the equity curve by more than 10 percent, a few worsened it, and most had a heavy profit penalty. Changing exit strategies often seems to degrade month-to-month performance. Hence, in the next section we will try to get a smoother equity curve by changing system design.
