The ACD Method

Discussion in 'Technical Analysis' started by sbrowne126, Jul 16, 2009.

  1. A small non-sequitur... but I'm pretty proud of this little histogram as it represents a decent amount of work. It's been said that ACD helps filter out the noise, and it appears I've found one way of doing that. This is over a three year period for this particular security.
     
    #13551     Jul 7, 2017
  2. Would you mind explaining more about your findings?

    Thanks.
     
    Last edited: Jul 7, 2017
    #13552     Jul 7, 2017
  3. Sure, the histogram is essentially showing the historical returns based on entering a trade given specific conditions. The x-axis is the 5 Day Number Line, y-axis being the % returns (should actually be the value*100 for the %), and the color bar is the 10 Day Number Line. There are two more filters applied, but as you can see these conditions have yielded a 100% historical hit rate.
     
    #13553     Jul 7, 2017
  4. Understood, nice chart.

    Out of curiosity is that plotted with ggplot2 using alpha fill? Looks a little R'ish to me :).
     
    #13554     Jul 7, 2017
  5. This is all python-driven research (pandas, numpy) with matplotlib being used for charting/graphs.
     
    #13555     Jul 7, 2017
    redbaron1981 likes this.
  6. Maverick74

    Maverick74

    In order to test this, you have to account for some alternative. In other words, had I randomly bought that same security over the 3 year period, what were the returns and are your results statistically significantly different then the null comparison. Otherwise the results are meaningless.
     
    #13556     Jul 7, 2017
    redbaron1981 likes this.
  7. Maverick74

    Maverick74

    Please guys, when you run tests, make sure you are testing your ideas against a null alternative. It doesn't even have to be the same stock, it can be the S&P 500 if you are testing stocks.

    Alternatively you can run stock XYZ over the performance of the ES and use the sharpe as the output.
     
    #13557     Jul 7, 2017
    koolaid likes this.
  8. Thanks for the suggestions, these are just preliminary data points as I'm not that far into it yet...
     
    #13558     Jul 7, 2017
  9. koolaid

    koolaid

    Huh? I thought you didn't care for ACD? Now you've been coding and doing data analysis for a system you don't care for?
     
    #13559     Jul 8, 2017
  10. I believe you're talking to DeltaRisk... Our handles are too similar!
     
    #13560     Jul 8, 2017