• Stock Selection System 20D versus 5D Rotation

    Since 20DMF signals might be kept active for long period of time, it is interesting to analyse the opportunity to rotate while a signal is still active.

    For example, the 20DMF was under a buy signal between November 26, 2012 and February 25, 2013.
    This is a three months period during which the S&P500 gained 5.8%.

    We can see below that if we had kept the original list of stocks purchased on November 26, it would have produced a return of 4.78%. A 20 Days rotation would have produced a return of 17.44%, while a 5 Days rotation would have produced a return of 20%.

    If we take into account the amount of extra risk and commissions linked to more frequent rotations, it seems logical to take 20 Days as a better rotation period.



    However, the two tables below show a different interpretation:

    The first table compares 5D to 20D rotation for Buy signals (which usually last for much longer periods than short signals.) We can see that during the Bull Phase that started from 2009, the 20 Days rotation period led to better returns. However, for the Bear market prior to March 2009, a 5 Days rotation would have been better.

    The explanation for such differences could be the following:
    - During a strong bull market, it is better to keep stocks for longer period so that we can catch the whole trend.
    - During a bear market, since the buy signals can quickly revert, a 5 Days rotation period forces to take profit earlier. It is similar to a passive money management method.

    For Short signals, it seems that the 5 Days rotation period always leads to better returns, simply because of the high volatility that characterizes short signals.



    In conclusion, more than the rotation period, what this analysis shows is that "Money Management" is an important issue: instead of rotating stocks on fixed periods, it is maybe best to exit losing trades and replace them by new stocks with better characteristics.

    I suspect that the 5D rotation period is a "poor way" to operate money management, even though in more volatile environments, it shows better results than a passive 20 Days rotation.

    We will therefore study money management in more detail.
    Comments 2 Comments
    1. mushin2003's Avatar
      Was there a follow-up on "...money management in more detail"?
    1. Pascal's Avatar
      Quote Originally Posted by mushin2003 View Post
      Was there a follow-up on "...money management in more detail"?
      Good question!


      I did not write a theoretical paper on money management, for two reasons: first there is not much to write and second, the "Portfolio Management" section is 90% focused on that subject.

      The back-tests have shown that in order to maximize the 5Days gains, the ideal stop was 5%. For 20 days gains, the stops must be raised to 8%.

      I have shown in the Portfolio Management that an active stop management allows for much smaller initial stops sometimes around 3%, if we can be patient enough and wait for a "good price."

      A good price is defined as a price that is closer to the good side of the buy range, the side that maximizes the profits and the Win ratio.

      The problem with that strategy is that it is sometimes difficult to find good trades early in a trend, such as for example for the short signal of yesterday: today, there is no acceptable trade. In that case, I do not hesitate to invest 50% in SDS or SSO and reduce that position as I find acceptable trades.



      Pascal