By David Vomund
In the April 1997 issue of the Opening Bell, we discussed a very simple but surprisingly effective strategy which combined AIQ’s market timing with the Relative Strength-Strong report. This strategy produced an average yearly return of 32.8%. We received more feedback from this article than any other this year, and a few subscribers were kind enough to send the results of additional testing that they performed. In this article we’ll report results from testing done by one of our subscribers and results of a few additional tests of our own.
Before we discuss the new testing results, let’s review the strategy that was used in our first study which tested the effectiveness of the Relative Strength-Strong report going back to mid-1992. The strategy was simple.
Feedback from our readers: testing confirms that the Relative Strength-Strong report works well...buying stocks with strong relative strength is a winning strategy.
Using a database of the stocks that comprise today’s S&P 500 index, we ran the Relative Strength report anytime there was an AIQ market timing buy signal (an Expert Rating of 95 or greater). Only the first signal in a string of buy signals was used and no confirmation was applied to the market timing signals. The top five stocks that appeared in the Long Term section of the report were purchased and held until a market sell signal was registered, at which time all positions were moved to cash.
To obtain the average yearly return of 32.8%, the following assumptions were made:
• The portfolio started with $25,000 in September 1992.
• All profits/losses were reinvested in the next block of trading with each of the five stocks receiving an equal amount of capital.
• The buy and sell points for the stocks were the opening prices the day after the market timing Expert Rating.
• Commissions were factored in using discount brokerage rates ($33 per stock trade).
• Slippage, dividends, and money market interest received were not factored in.
After performing this test, several questions remained:
Would results be higher using the report's Short Term section rather than the Long Term? How would the results differ if the strategy was used on a larger database instead of S&P 500 stocks? How would this strategy fare under poor market conditions? What if more stocks were purchased instead of just five?
Of the feedback that we received, the most extensive testing came from Dennis Nichols. Mr. Nichols posted his results on his Trading Connections Internet home page at www.tradingconnections.com. (Trading Connections develops add-on programs for TradingExpert.)
Here are a few of Mr. Nichols’ findings:
Study #1, Eight-year Test
Mr. Nichols ran the same test as we did (see above) but his was an eight year test beginning on 11/09/88. Buying the top five stocks from the Long Term section of the Relative Strength-Strong report produced a compounded annual rate of return of 42.9%. He appropriately cautions readers that this was based on today’s S&P 500 structure so some of the stocks that were purchased may not have been in the structure at that time.
Study #2, Broader Database
Mr. Nichols ran the same eight year test using a broader database of 1600 stocks that have been traded since 1988. By purchasing the top four stocks on the report, the return was 41.8%. He notes that this is somewhat overstated because the stocks that have been trading since 1988 have passed a "survival of the fittest" test.
Study #3, Using Short Term Section
Mr. Nichols’ next test used the Short Term section of the Relative Strength-Strong report. By purchasing the top three stocks on the report using a database of the current S&P 500 industry group structure, the trading strategy produced a return of 42.3%. This result is very similar to his test which purchased the top five stocks on the Long Term section of the report.
Study #4, Effect of Market Timing
Our simulated testing included two elements: market timing and stock selection. Mr. Nichols wanted to test the effect that market timing had on the simulated results.
Using the same number of trades as AIQ’s market timing signals, a series of randomly timed trades was created for the 11/09/88 to 01/06/97 testing time period. He found that buying the top five stocks from the Relative Strength-Strong report using S&P 500 stocks averaged 31.7% per year. If you purchased the S&P 500 index rather than high Relative Strength stocks, your annual return would only have been 9.4%.
Many thanks to Dennis Nichols for performing this research and posting the results on his web site. Now for some of our own studies.
Results of Purchasing Ten, Instead of Five, Stocks
In our work published in the April Opening Bell, we always used the top five stocks on the Relative Strength report. To only hold five stocks, however, is risky and will lead to a volatile portfolio. Our first test is to see how the results change if ten stocks were purchased after each market timing buy signal instead of five stocks. In this test, we looked at the performance of the sixth to the tenth highest rated stocks.
Table 1 shows the trade by trade details of this test. We would expect that the top five stocks on the report would outperform the next five stocks and that is what happened. The top five stocks on the report had an average return per trade of 9.04% while the next five stocks had an average return per trade of 6.07%. While this is a lower value, it is still well ahead of the S&P 500’s 2.7% return per trade.
Table 1 Long Term Relative Strength Strong Report Details Stocks 6 to 10 E.R. Buy E.R. Sell S&P 500 Rel. Str. Signal Signal (% ch) (% ch) Stocks Held (ticker symbols) 09/10/92 09/22/92 -0.67 -2.08 FTL,U,WWY,AMD,SFA 09/28/92 12/14/92 3.89 31.17 COMS,DELL,WIN,UIS,SFA 12/18/92 01/07/93 -2.39 -0.98 SFA,WIN,BAY,CC,BKB 01/12/93 02/16/93 0.67 -5.15 CSCO,BAY,C,SFA,AMD 07/06/93 10/22/93 4.95 33.54 SLI,TLAB,BAY,GIC,ORCL 12/17/93 02/04/94 0.74 3.59 PDG,EMC,HFS,ORCL,DIGI 02/28/94 03/24/94 -0.05 0.54 COMS,HUM,HFS,ANDW,HLT 03/28/94 06/20/94 -0.99 -8.73 HUM,COMS,CPQ,THC,ECO 06/27/94 08/05/94 2.17 7.84 COMS,LSI,AR,MIL,CEN 08/23/94 09/19/94 1.36 6.77 ANDW,DELL,TLAB,TDM,LOW 09/26/94 09/29/94 0.31 3.45 AR,HUM,TEK,UK,ANDW 10/10/94 10/20/94 1.67 0.23 ANDW,HUM,AMH,MU,ADSK 11/07/94 04/20/95 9.19 18.15 DELL,TEK,TDM,MU,CPQ 04/21/95 06/16/95 6.16 -2.60 MDT,EC,AMH,TEK,BMET 08/25/95 10/02/95 3.86 4.04 DIGI,TLAB,DELL,COMS,CSE 10/12/95 10/19/95 1.29 2.44 BAY,CSCO,LSI,MDP,GNT 11/16/95 12/18/95 1.59 -11.14 CSCO,CS,BSX,DELL,COMS 01/16/96 04/03/96 7.80 23.09 PX,S,NEM,TMC,U 04/15/96 04/17/96 -0.14 0.85 Z,SUNW,ANDW,CSCO,U 05/08/96 06/07/96 4.32 7.43 CSCO,USS,SUNW,RDC,SMED 07/09/96 08/29/96 0.41 7.70 CSCO,GPS,CHRS,WCOM,DELL 09/09/96 01/06/97 12.64 13.43 RDC,HFS,NKE,HLT,CPB Average = 2.67 6.07
Looking at yearly returns, the April Opening Bell article reported that trading the top five stocks on the report from 1993 through 1996 led to an average yearly return of 29.2%. Using the next five highest rated stocks, (stocks six to ten) the average yearly return drops to 19.4%. Therefore, using this trading strategy on the top ten stocks would lead to a return of about 24% per year. This compares to the S&P 500’s 15% yearly return. The overall return is lower when 10 stocks are purchased instead of 5 but the volatility will be lower as well.
Results of Test Using Short Term Section
In our next test, we looked at the results from using the Short Term section of the Relative Strength-Strong report. We ran the exact same study as reported in the April Opening Bell but purchased the top five S&P 500 stocks that appeared on the Short Term section of the report. Whereas the Long Term section of the report looks at a one year time horizon, the Short Term section looks at a six month time horizon. The trade-by-trade results are found in Table 2.
I was surprised how many of the same stocks appeared on both the Short Term and the Long Term section. As a result, the overall percentage return per trade is similar. The average trade using the Long Term section gained 9.04% while the average gain per trade using the Short Term section was 7.07%.
Table 2 Short Term Relative Strength Strong Report Details E.R. Buy E.R. Sell S&P 500 Rel. Str. Signal Signal (% ch) (% ch) Stocks Held (ticker symbols) 09/10/92 09/22/92 -0.67 -0.65 DIGI,ADSK,WWY,SFA,HAL 09/28/92 12/14/92 3.89 20.56 DIGI,COMS,WIN,RDC,WWY 12/18/92 01/07/93 -2.39 8.58 DIGI,COMS,DELL,BAY,EMC 01/12/93 02/16/93 0.67 -7.79 DIGI,COMS,DELL,HET,INTC 07/06/93 10/22/93 4.95 21.53 ECO,DIGI,MU,HM,EMC 12/17/93 02/04/94 0.74 5.53 COMS,TLAB,AMH,THC,HET 02/28/94 03/24/94 -0.05 12.67 COMS,STO,CPQ,THC,BAY 03/28/94 06/20/94 -0.99 -5.99 MU,DELL,STO,AMD,COMS 06/27/94 08/05/94 2.17 9.87 MU,LSI,AMD,ANDW,NUE 08/23/94 09/19/94 1.36 5.61 SFA,AMGN,MSFT,TLAB,IBM 09/26/94 09/29/94 0.31 0.85 LSI,TLAB,AMH,COMS,SFA 10/10/94 10/20/94 1.67 1.19 LSI,CA,DELL,USS,VIA-B 11/07/94 04/20/95 9.19 16.99 LSI,DELL,AMH,CSCO,AAPL 04/21/95 06/16/95 6.16 28.15 EC,U,MU,DEC,BMET 08/25/95 10/02/95 3.86 3.91 MU,AMAT,SUNW,LSI,CSCO 10/12/95 10/19/95 1.29 4.82 U,HFS,BAY,CSCO,MDT 11/16/95 12/18/95 1.59 -6.78 SUNW,HFS,BAY,DGN,CSCO 01/16/96 04/03/96 7.80 20.40 HFS,RDC,SUNW,DEC,NKE 04/15/96 04/17/96 -0.14 -0.89 CHRS,RDC,GAP,SMED,Z 05/08/96 06/07/96 4.32 10.45 CHRS,ENS,TDM,RDC,TJX 07/09/96 08/29/96 0.41 -6.17 CHRS,Z,RDC,KM,ANDW 09/09/96 01/06/97 12.64 12.71 DELL,BG,NAE,CHRS,ORX Average = 2.67 7.07
The overall return was comparable but the results were less consistent. In 1994, trading results using the Long Term section was 28.9% but the Short Term section only produced a return of 2.3%.
We again caution that these results are overstated because they are using the stocks from today’s S&P 500 structure. The only way to get a true test is to use the S&P 500 structure that was available at the time of each market timing buy signal. That is beyond the scope of this article. What we did instead was to create an industry group that contains all of today’s S&P 500 stocks. Since this is the database of the stocks we are trading, it makes sense to compare our trading results to this S&P 500 industry group.
We found that the industry group that contains all of today’s S&P 500 stocks did outperform the S&P 500 index, especially when you go back to 1993 where many of the stocks in today’s S&P 500 index were not a part of the S&P 500 structure at that time. The degree of outperformance over the entire testing time period was not significant, however. From 1993 through 1996, the S&P 500 index averaged 14.97% per year. The average return on the group of stocks that we used for our studies gained 15.90%.
The results from all of the tests we performed are found in Table 3. The best return was seen by using the Long Term section of the Relative Strength-Strong report with AIQ’s
market timing model.
How Does the Strategy Work in Bear Markets?
In our final test, we looked at how this stock picking strategy works in grim market environments. Since no market timing model works all the time, we want to see how this stock selection system works during corrections/bear markets. The best trading strategies outperform in both bull and bear markets.
In this final test, we looked at four of the worst time periods to be long in stocks in the last seven years. We assumed that you bought the day of the high and sold on the day of the low. Therefore, the Relative Strength report was generated at the market’s high point and the top five stocks on the Long Term section were purchased and held until the low was made in the market.
Table 4 shows the time periods used and the results from trading the stocks based on the report.
The first time period listed is the bear market of 1990. From July to October, the S&P 500 index fell nearly 20%. An unweighted index of today’s S&P 500 stocks fell 24.8% over the same time period. If you ran the Relative Strength report and purchased the top five stocks, you would have lost 27%.
Another bad period was the short but sharp drop in the summer of 1996. The S&P 500 fell 7.25% but the five high Relative Strength stocks fell an average of 15.4%.
During our recent March/April correction, the Relative Strength stocks fell about the same amount as did the market.
The best time period was in early 1994 when the five Relative Strength stocks rose at the same time that the S&P 500 fell almost 9%. This was due to one stock, Micron Technology, which rose 43% over that time period. Unfortunately, Micron Technology was not a part of the S&P 500 at that time.
From our testing it appears that this trading strategy will typically underperform during bearish time periods.
The results of our Relative Strength report testing are encouraging for those like myself who have been preaching the benefits of buying into strength rather than trying to call the lows. I’m not saying that counter-trend trading is ineffective — many people are able to call the low points. Rather, unless you are very good at identifying support levels and reading the indicators, buying into strength is an easier and, for most, a more profitable strategy.
We ended our current studies by looking at bearish time periods, which is important. Quite often, high Relative Strength investors outperform during bullish time periods and underperform during bearish time periods.
Anyone applying a high Relative Strength strategy must be willing to accept volatility. There are ways to lower volatility. Market timing is one of them. Performing some fundamental analysis on the high Relative Strength stocks is another method (see May 1997 issue of the Opening Bell).
Next month we’ll look at multiple technical screenings run on high Relative Strength stocks.Back to hints and tips