• When to buy the weakest or the strongest stocks on a buy signal

     
    4 Comments Comments
    The 20DMF produces different types of buy signal, depending on the level of the OverBought OverSold indicator:

    - Buy Oversold: When the OBOS < -70
    - Buy Protected 70 when the OBOS crosses over -70
    - Buy Protected 30 when the OBOS crosses over +30
    - Buy when the MF crosses over 0%

    Past studies have shown that when a Buy Oversold signal is produced, it is better to buy the stocks from the weakest sector and among them select the stocks with the weakest price relative strength (Price RS.)

    However, on a standard Buy signal, it is better to buy stocks that show a strong price RS and belong to a strong sector.

    The question is: when do we transition from buying the weakest to buying the strongest stocks? Since buying weak stocks can ride a wave of short covering, we can conclude that the Oversold character of the market should determine the type of stock to buy.

    The tables below show the summary of a stock selection analysis that depends on the level of the OB/OS indicator of the 20DMF. This analysis is one element of a stock selection process that we plan to start soon.

    The first Table shows the results from buying the strongest stocks, while the second table shows the results from buying the weakest stocks. Note that the stock selection process is reset after 5 days if the 20DMF Buy signal stays valid. This means that every 5 days, we search for the stocks with the best settings.

    There are four levels of OB/OS: Below -50, between -50 and 0, between 0 and +80, Above +80.

    The combination of the two tables shows that:

    - Below -50, it is better to buy the weakest stocks, because the market is in Oversold mode.
    - Between -50 and 0, it is difficult to make money. Buy signals are weak and prone to failure. The market is not in oversold mode anymore and not yet in an uptrend. Stay in Cash.
    - All the strong uptrends occur when the OBOS is between 0 and +80. The tables show that buying the strongest stocks leads to better returns.
    - Above +80, the overbought character of the market becomes evident and it makes sense to short the strongest stocks (even during a buy signal.)

    A few other points on the Tables below:

    1. Most trades occur either in Oversold mode or during a strong uptrend.
    2. Even though the Oversold mode trades offer a better return, the sheer number of trades that occur during a strong uptrend dictate that most of the portfolio gain will take place when buying leading stocks.
    3. The gains are calculated on 5 opened days. 1% gain over 5 days means that in theory, the yearly return would be 1% * 220 / 5 days = 44%
    4. We can see that this stock-picking strategy yields more than twice the S&P500 returns over the same period.



    Comments 4 Comments
    1. xr-3609's Avatar
      Pascal,

      Are the "Best stocks" & "Worst stocks" in the Overbought/Oversold diagram immediately above the same stocks that are found in the "Best Stocks" & "Worst Stocks" filters?

      I assume there is a difference because you say above:

      "Past studies have shown that when a Buy Oversold signal is produced, it is better to buy the stocks from the weakest sector and among them select the stocks with the weakest price relative strength (Price RS.)

      However, on a standard Buy signal, it is better to buy stocks that show a strong price RS and belong to a strong sector."

      Which would seem indicate for example that the worst stocks are selected solely on the basis of worst 20D Price change from worst sectors as below?

      SECTORS 20D Price change
      Coal -18.6%
      Banks Foreign -11.6%
      Internet Networking -10.2%
      Home Builders -9.9%
      Silver -9.5%
      Base Metals -9.4%
      Gold -9.4%
      Cloud Computing -8.7%
      Copper -8.5%

      Thanks,
      Greg
      Reply
    1. Pascal's Avatar
      Quote Originally Posted by xr-3609 View Post
      Pascal,

      Are the "Best stocks" & "Worst stocks" in the Overbought/Oversold diagram immediately above the same stocks that are found in the "Best Stocks" & "Worst Stocks" filters?

      I assume there is a difference because you say above:

      "Past studies have shown that when a Buy Oversold signal is produced, it is better to buy the stocks from the weakest sector and among them select the stocks with the weakest price relative strength (Price RS.)

      However, on a standard Buy signal, it is better to buy stocks that show a strong price RS and belong to a strong sector."

      Which would seem indicate for example that the worst stocks are selected solely on the basis of worst 20D Price change from worst sectors as below?

      SECTORS 20D Price change
      Coal -18.6%
      Banks Foreign -11.6%
      Internet Networking -10.2%
      Home Builders -9.9%
      Silver -9.5%
      Base Metals -9.4%
      Gold -9.4%
      Cloud Computing -8.7%
      Copper -8.5%

      Thanks,
      Greg
      You are right Greg, my explanations regarding the stock selection process itself is not too good.
      I will republish here a post that I originally made on the Model section. I will rename it:
      "Stock Selection System results without rotation"

      This post explains the stock selection process.
      It is somewhat different from the filters, which were developed two years ago.
      For example, two years ago, I had not developed the "number of days" until a buy/short signal that I calculate on a sector base. I use this process to find the best/worst sectors simply by ranking these sectors using the distance to their Buy/Short signal.

      If you look at the PascalA list (First link in the Filters section,) the Sectors Sheet indicates for each sector his distance to a Buy/Short signal. This is what I have been using to rank sectors.

      Attachment 19001

      On a Buy Oversold signal, the "Best stocks" filter points you to the stocks in the worst price RS section

      Attachment 19000

      And in that section, the "Best stocks" filter points to the stocks that have the strongest Thrust

      Attachment 19002

      This "Best Stocks" Filter works very nicely on 20DMF Buy Oversold signals.
      I back-tested it some years ago and have used it since then.

      The new stocks selection system we are developing now uses more recent tools, but will also be able to pull the nicest stocks before the market closes depending on the Real-Time situation of the various indicators.

      One important aspect also of the most recent back-test is that we now have much more data than what we had two years ago. For example, we are using the June 2011-June 2013 period to set the process and then we test the process for the bear market of Mid 2007 to March 2009 and the Bull market of March 2009 to June 2011.
      Reply
    1. frank's Avatar
      Hi Pascal,

      This is very interesting study which deals with what to buy when the ob/os signal is rising from bottom -100 to top 100 levels.

      Is there a similar study done for what to sell (i.e. for sell signals) when the ob/os signal is dropping from top 100 to bottom -100 levels?

      Thanks,
      Frank
      Reply
    1. Pascal's Avatar
      Quote Originally Posted by frank View Post
      Hi Pascal,

      This is very interesting study which deals with what to buy when the ob/os signal is rising from bottom -100 to top 100 levels.

      Is there a similar study done for what to sell (i.e. for sell signals) when the ob/os signal is dropping from top 100 to bottom -100 levels?

      Thanks,
      Frank
      I did not find a way to rotate from one type of stock to another type of stock during a pull-back, because in general pull-backs have been swift and reversals have occurred faster than the time it takes to rotate from one weak sector to another weak sector.

      In general, when there is a short signal, it is better to short the stocks with the worst price relative strength, because these are the stocks that funds will liquidate first (they try to keep their strong stocks or will even add.)



      Pascal
      Reply