ASRT - My Methods For Finding The Best Stock Portfolio Part 3

2023-05-26 11:10:38 ET

Summary

• Questions and comments from readers answered.
• Lookback timeframes identified.
• Selection of original list of quality stocks from which to select best.
• Technique of buying and selling after the best stocks are identified.

I hadn't originally planned to write another part of this article, but I felt compelled to do so because of the questions submitted by readers of Seeking Alpha. I am very glad they asked those questions, as I was forced to do the intensive research I had to do, which I had previously avoided because I knew how much work it would be. As it turned out, I underestimated how much work it would be and how much time it would take to do the study, but having done it at this point, I am glad I did.

The main question presented was why did I use October 3rd to begin the lookback period rather than 3 or 5 years back as most analysts do? Did I think that this was far enough back to isolate the strongest trends in stock prices? I selected that period because I reasoned that things might change for some unknown reason in the market at changes in general trend, and I wished to pick up on whatever new trend there might be. It turned out that I was right for the reasons I explained in Part 1 . I continued to explain my methods in Part 2 .

The main question I intend to answer in this part is the timeframe(s) I use to find the stocks with the greatest potential to make the best gains over the next few months. The first question is how far in the future can we expect our results to apply? The study I did was to look at the lookback period, but it showed me that the ideal forward timeframe was also important too. I was hoping for 6 months in the future but found that the best time was consistently one month in the future, though the stocks seemed to continue to excel for much longer. I was impressed but not shocked at this finding. It is logical because trends seem to change often everywhere else so why not in the market as well? If one assumes that a trend will last only a month (or less), then if a stock remains on the list for 6 months or longer, it is a safe bet that it will continue to produce growth for a longer and longer time. As I said before, Nvidia (NVDA) has been on my list for years and continues to grow. The events of this last Thursday justify my decision to hold.

The Time Frames

Now for the meat. I tried several lookback timeframes and combinations of timeframes to best estimate which ones are best. Knowing that I was not the first one to do this, I took leads from others. I remember several studies I had done in the 90's, and those reported by Mark Boucher and Tim Hayes. Each of us used a combination of timeframes of varying lengths. Mark Boucher used up to 11, and Tim Hayes used 3 as did I. I tried up to 12 and as little as one. The results were as follows:

1. For one look-back timeframe, I found that 85 days (17 weeks) was the best single look-back timeframe to use. This is an average of all my test results using a one-month forward result. I began using 1 to 6 months ahead but soon saw that the results over 1 month failed to show significant consistency. The range was between 69 and 100 days.
2. For a 3-period lookback, I found that a combination of 30, 75, and 194-day lookback were the best averages to use. I did this by doing a Rate of Change for each day and averaging them out. In other words, I used the current price (for that day) divided by the price 30 market days ago (6 weeks before), plus the current price divided by the price 75 days ago (15 weeks prior), plus the current price divided by the price 194 days ago (39 weeks prior), and divided the result by 3. I did this for all the stocks in my group, then ranked them by the top results, and averaged the top 9 stocks to get a result. I repeated this process for various dates from 2011 through the present. I used a program called "Genehunter" by Ward Systems to do the heavy lifting of finding the best combinations. I've been using this software since 1994 and find it very useful for all studies of this type.

I tried to develop systems using other combinations of timeframes but found that there were too many problems with using more than 3. The range of the different look-back periods was too extensive. I found that 3 timeframes were sufficient to use for this purpose.

Selection of Stocks

Another comment was how I selected the group of stocks to use as a base list. As I said in Part 1 , I piggybacked from established systems of value-oriented stock screeners by other analysts. Ultimately, why should one stop there? There is no reason that one could not use a growing set of stocks from any source. If you think that a stock is valuable, add it to your list. Who says that you couldn't have a set of 200-300 stocks you think are good enough to add? Maybe start from a list of all stocks with an RS of 80 or more, or a list of the stocks recommended by the WSJ? It is up to the individual trader. If the guidelines are followed (top 9 stocks or any number you feel comfortable with, and no shooting stars), the sky is the limit. More than likely you would come up with a similar group as me, or close to it. The list could also be revised whenever you feel that some of the stocks have lost their value to the overall selection. Add others any time you think there is potential. The only real problem I found is that the larger the list, the longer it takes to keep them up to date. I use Excel to do this analysis and it does a good job, but I have to manually update the data whenever a stock pays a dividend or splits.

As far as selling your stocks when the time comes, I use my experience and technical analysis to tell me when that time is at hand. I find that even good stocks need to be dumped if they suffer an unusual decline that shouldn't have happened. I find that trend lines (support and resistance) are the most efficient tools for this process. The trick with trend lines is that if the stock violates one and then proceeds sideways, a false signal will occur. This is where experience comes into play. It's always better to exit than to lose in a significant decline. You could always buy it back if you think it has regained its strength.

Conclusions

In attempting to isolate the ideal set of stocks in which to invest, the leaders of the trend in the overall stocks in the market change often. In order to keep up with the changes in trend, it is necessary to isolate those with the greatest potential to grow now, if that is your goal. It may not be if you are relying on dividend-paying stocks or some other criteria. Even for the investors that desire long-term gains, this method would indicate which stocks are likely to continue to provide gains in the long term or help to understand which stocks to avoid when compared to others in its group. The purpose of this study was to identify those stock issues that have the current support to make gains in contrast to others in a group of like stocks.

The recommended ROC periods are 6 weeks, 15 weeks, and 39 weeks, averaged out and compared to all others. This system can be used for any group, sector, or industry to identify the best candidates within that group.

Even though this study has been extensive, it is not absolute. The numbers are averages within a range so results could vary from actual final results. Fortunately, the range of the 3-period ROC group was very narrow, making me very confident with this result. Use your experience and technical studies to make decisions about when to buy and sell. This system only identifies the best stocks in which to invest at the current time and is not a "buy and hold forever" system. This is why I used so many value-based systems to identify good long-term high-probability investment opportunities.

My current (as of May 24, 2023) list is the following:

VKTX , SMCI , MLTX , OPRA , RIOT , NVDA , ASRT , LMB , BLDR , ELF , AMD , ETNB , SHOP , MSFT , & FSLR

This list is based on the 85-day single look-back method. I intend to upgrade to the 3 ROC system soon which may change things a little, though so far, my studies show little change in the list between systems. My recommended list is the first 9 stocks and the final 6 stocks are for those who want to expand their diversification. There is a dispute among advisors as to how many stocks should be in a good portfolio. I have heard anywhere from 10 to 50, but in my opinion, 9 is pushing it. My ideal portfolio is between 5 and 7 different issues. I went with 9 to ensure that the portfolio is diversified and remains consistent for a longer period. This system is self-correcting in that if a stock in the top 9 falls below position 20 (or whatever level you want), it is eliminated and replaced by one that becomes elevated into the top 9.

For further details see: