The Long-Term Relative Strength-Strong Report with Market Timing

David Vomund, writing in the April 1997 issue of AIQ Opening Bell, presents a mechanical trading strategy using the AIQ TradingExpert for Windows Relative Strength-Strong report and other reports (the text and tables of David's article are available online). The simulated results of using the Long Term Relative Strength-Strong report along with the AIQ market timing signals was impressive, producing an average annual gain of approximately 3 times that of the S&P 500 index. The systems using other reports that he presents in his article did not perform well, so the rest of this study is concerned only with the Relative Strength-Strong report.

What David did

David used an account of $25,000 starting with the buy signal on 09/10/92, buying equal dollar amounts of the top five stocks from the S&P 500 on the Long-Term Relative Strength-Strong report for the day of the market timing signal. The purchases were made at the opening price of the day following the market buy signal and sold at the opening price on the day following the market sell signal. Other conditions included: signals were market ERs of 95 or greater; commissions of $33/stock trade; no allowance for slippage, dividends, or interest when out of the market.

What we wanted to know

We wanted to know how well this method worked in the period preceding David's starting date of 09/10/92, and how well the method worked for a stock universe larger than the S&P 500. We were also curious about using stops or minimum volume criteria to improve the performance of the system.

What we did

We did two studies, both starting with the market buy signal that occurred on 11/09/88. One study used only the stocks from the S&P 500, while the other study used a collection of 1600 stocks representing (mostly) stocks that were trading at a volume of at least 10-20,000 shares per day a few years ago and which were still trading today. A minimum price of $5.00 was required for this larger universe (a not entirely satisfactory restriction, but that's what we did). Note the same caution that David made - the makeup of the S&P 500 that we used was current as of 03/01/97 while the S&P 500 of 01/20/88, and times in between, certainly had some different components.

Using Dial/Data market data, there are 22 periods when the AIQ market signal had us in the market between 11/09/88 and 09/08/92 (the time period before David's test period), and also 22 in-the-market periods from 09/10/92 to 01/06/97.

What we found

This graph shows our findings most clearly. Note that the horizontal axis is not to scale.

Specifically we found::

• Buying the top 4 stocks from the Long-Term Relative Strength-Strong report on the larger universe of 1600 stocks gave the optimum return for that universe, averaging 41.8% per year compounded for the 8+ years studied.
• Buying the top 5 stocks from the Long-Term Relative Strength-Strong report on the S&P 500 universe gave the optimum return for that universe, averaging 42.9% per year compounded for the 8+ years studied.
• The system's return on the larger universe of stocks was very good from early 1989 to early 1994
• The system's return on the S&P 500 universe of stocks was very good from late 1990 to the end of the test
• At 14.4% per year, the timed S&P 500 index barely out-performs a Buy and Hold strategy at 13.1% per year.
• Almost any setting for principal protection stops or minimum volumes gave lower returns

What does this mean to you?

This is not a recommendation to use this system for trading or to make any specific investments. It is only a study of simulated returns on historical data. The results with the 1600 stock universe, while probably valid, are not possible to duplicate due to the unique nature of the particular collection of stocks. The results for the S&P 500 should be able to be duplicated. Interestingly, our results were not exactly the same as David's. Some of the time, only 4 of the 5 strongest picks were the same, with the 5th one differing for some undetermined reason(s). The simulation results were, however, quite close.

Now what?

We did a control study of sorts and obtained very interesting results. In addition, one of our AIQ-using friends had another idea, so we did a study on the Short-Term Relative Strength-Strong report, too. You should take a look at both of them. The is also a Summary page that brings many of the results together.

Home AIQ Software Contacts Studies