1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

EA backtesting optimization and back-tested results with different parameters? Please

Discussion in 'Forex Trading Systems and Strategies' started by raysum, Oct 30, 2011.

  1. raysum

    raysum Recruit

    Joined:
    Aug 22, 2011
    Messages:
    1
    Likes Received:
    0
    hi everybady, I encountered a problem, I was doing optimization, the optimization of the parameters out, and then re-set period of time once and optimize the same outcome and optimization of the results are not the same, you have been such a situation?

    In the first picture,For example, I now do a history of EA optimization time is 2011/6/17 ~ 2011/10/17

    2 pictures, I chose the best parameters in this group

    Image 3, when I put a good optimization parameters, the test once again historical information, at the same time

    4 pictures, test out and optimize its outcome is completely different, why is there such a big difference?

    5 pictures
    And optimize the results do not like the result of optimization is profitable 61.34 yuan, Total Trades: 79, Profit Factor 1.07, Expected Payoff: 0.78, using the optimized parameters back to the measured result is Trades: 22, Profit Factor 0.62, Expected Payoff: -6.85, why is there such a big difference? I also encountered another problem, good parameter to adjust a few days ago, in a few days once again test the same time, the result is completely different, some people come across this problem?
     

    Attached Files:

  2. Pharaoh

    Pharaoh Colonel

    Joined:
    Oct 3, 2007
    Messages:
    19,657
    Likes Received:
    2,205
    The 2 lines on the graph in Image 4 aren't identical, but do follow roughly the same pattern.

    Could you please clarify the difference between the two tests?
     
  3. TadS

    TadS Private

    Joined:
    Apr 3, 2013
    Messages:
    21
    Likes Received:
    0
    When your results vary greatly after altering a few parameters, it might be a sign of an unstable system. An unstable system will vary greatly with changes in market conditions, parameter settings, or anything else that might affect the system.

    You should be very careful moving forward with an unstable system and may want to revisit the fundamentals values behind the strategy.
     
  4. Wish6549

    Wish6549 Recruit

    Joined:
    Nov 1, 2013
    Messages:
    3
    Likes Received:
    0
    If you have to spend ages back testing and fine tuning your EA it means it isn't going to work when you use it in the real work.
     

Share This Page