Probability Meter

Hey guys,

How about creating a probability meter for single currency pair (or other instrument). It's like a strength meter, but instead of comparing and weighting a ton of currency pairs against each other (strength meter) this would be comparison of various indicators for same instrument within current and neighboring time frames to estimate probability of the continuation of a trend. 

Detailed explanation of how such an indicator would be built can be found here:

The difference I see would be representing that as a line indicator (if that's possible for BloodHound to print line subgraph) or traditional bar graph. The advantage would be that we had an overview on behavior of this indicator over past time and we could further optimize/analyze it as what probability percantage is likely to give us good possible entry signal or direction of the trend.

Shall we try it?

-Bart S.

Have more questions? Submit a request


  • 0
    Bart Slowinski

    Hey again,

    Here is my first attempt to build the Probability Meter as described above. 

    The first and foremost limitation I found is lack of indicators to match same list as per the original project that was done for MT4 platform. On the list of my solvers I marked them with the dash line '-' right behind the number.

    Obviously the total output will not be displayed in percent manner as in original, but in decimal equivalent.

    The other problem I encountered is how to tell solvers that moving average is increasing or decreasing? How to compare current value to previous one?

    That's it for now. See ya later.

    -Bart S.

  • 0

    @ Bart -

    I am the co-author of the nitro+ indicator referred as a general guide.  Don't mean to intrude.  My wife sent this link to me as she was scanning for DMCA candidates and stumbled here.

    For MAs, or any indicator that outputs to x-asis including MACD-type histograms, use 'SHIFT' in the code.  If that is possible.  Or maybe you knew this, and were just asking how to add SHIFT.  Forgive me if this is the case.

    For Example using generic logic (indicator, indicator period, method, symbol, periodicity, shift number of bars to right);

    ma_main=  (MA,  50,  LWMA,    GOOG,  1440,   0.0);

    ma_signal=(MA,  50,  LWMA,    GOOG,  1440,   1.0);

    if ma_main >= ma_signal,

          Your_LONG_MA_Parameter =   1.0 point;

          Your_SHORT_MA_Parameter = 0.0 point;


                  if ma_main  <  ma_signal,

                  Your_LONG_MA_Parameter =    0.0 point;    

                  Your_SHORT_MA_Parameter = 1.0 point;

    Yes, ">=" presents a longside bias issue.  With over 100+ parameters such as these across 4.7 TFs, we easily resolve it by making other binaries "<=".  The point system above of course does not matter. 

    Thanks, Steven


Please sign in to leave a comment.
Powered by Zendesk