Equity curves for filtered and unfiltered CHBO systems for DM

-FiltCHBO i -NO FILTER |

Date (2/88-6/95)

Figure 6.18 Equity curve for a 20-day channel system on DM with (upper line) and without (lower line) filter. Calculations are with rollovers on actual daily con­tract data.

202 Equity Curve Analysis

Filtering also produced an equity curve (see Figure 6.19) that is smoother than for the unfiltered 65sma-3cc system, described in this chapter as case 1. Note that the case 1 equity curve has been converted from daily to monthly data. The SE for case 1 was $3,776, versus $2,507 for the filtered equity curve, a difference of 50 percent. The interval drawdowns were also smaller, confirming the smoother equity changes (see Figure 6.20).

We then compared the performance of the channel breakout sys­tem with and without filtering. The average interval equity change over 1, 3, 6, and 12 months was greater for the filtered system, as shown in Figure 6.21, page 204. Thus, the filtered system produced more con­sistent results for DM. For example, standard deviation of interval re­turns was greater for the unfiltered system (see Figure 6.22, page 205), confirming its uneven performance.

As expected, a linear regression analysis showed the standard error for the filtered system to be $2,507 versus $3,937 for the unfiltered sys­tem. Thus, filtering produced a smoother equity curve, since the stand- Equity curves of 65sma-3cc and filtered channel breakout

•FiltCHBO -65sma3cc

Time (months) 2/88-6/95

Figure 6.19 The filtered breakout system (upper line) had a smoother curve compared to the 65sma-3cc system without initial stop or exits (case 1; lower line).

Effect of Filtering on the Equity Curve 203