Following the Trend

irom the previous chapters we have seen that short-term trading can be incredibly rewarding, but also very risky, especially if the market you are trading simply does not produce the moves you seek. Because not every market can produce a short-term average profit big enough in dollar terms, the only way to trade these markets is to go with less frequent trading and longer trades that raise the value of the average trade to a level worth trading.

In the previous chapters we discussed how to classify trades and what might constitute long term and short term. It might be a good idea to always try to be as short term as you can when it comes to market time units (bars). When you're designing a long-term system, measured in regular time units (days), you should use weekly (or perhaps even monthly) data during the building and research process so that you can work with as few market time units (bars) as possible. After an optimal number of time units for the system in question, the further away from the signal you move, the less profitable and reliable your results will be, as can be seen from Table 7.1.

Table 7.1 shows the average profit per trade from a 20-day (four-week) breakout system, traded on the long side only on the S&P 500 stock index futures contract over the period June 1983 to August 1999, and with all trades lasting for a certain amount of days (weeks) no matter what. From the top half of this table you can see that, in this case, the longer the holding period, the higher the average profit. But with a higher average profit comes a higher standard deviation of the individual outcomes, indicating that the uncertainty about the results also is increasing. This can be measured by the ratio between the average profit and the standard deviation. As long as the ratio is increasing as well, however, the average profit is growing faster

TABLE 7.1

Trade/time reliability comparison.

TABLE 7.1

Trade/time reliability comparison.

Days in trade

10

20

30

40

50

60

Avg. perc. profit

0.152

0.462

0.854

0.886

1.293

1.304

St. Dev.

2.033

2.822

3.368

3.842

4.913

4.982

Ratio

0.075

0.164

0.254

0.231

0.263

0.262

Profit/day

0.015

0.023

0.028

0.022

0.026

0.022

Weeks in trade

2

4

6

8

10

12

Avg. perc. profit

0.366

0.818

1.086

1.205

1.468

1.780

St. Dev.

2.136

2.651

3.433

3.559

4.571

4.746

Ratio

0.171

0.309

0.316

0.339

0.321

0.375

Profit/week

0.183

0.205

0.181

0.151

0.147

0.148

Profit/day

0.037

0.041

0.036

0.030

0.029

0.030

than the standard deviation, indicating a decreasing relative risk. In this case, the ratio seems to level off after about 50 days, indicating that there is little to be gained by staying any longer in the trade. This is confirmed by the average profit made per day, which starts to decrease after 30 days. Similar observations also can be made from the bottom, weekly part of the table. What is even more important is that by comparing the weekly data with the corresponding daily data, we can see that the weekly data produce both more profitable and less risky results, as measured by the average profit and the standard deviation. That said, in the following long-term systems we work both with weekly and daily data.

0 0

Post a comment