Possible to code this 3D heatmap in PRT? (Mark Whistler’s Wave PM)

• 06/28/2019 at 8:50 AM #101586
  Nicolas

  I’m going around in circles and I’m convinced that the problem comes from the platform engine for calculating values in a loop.

  attached 2 screenshots explaining the phenomenon. The code below is the heatmap indicator that records the value of each indicator for each calculated period. I have voluntarily painted in red a period value to check (the period 200), applying the indicator “wave PM” on the graph with the same period can thus control the values returned by the loop.
  We see that nothing matches. The “wave PM” indicator marks extremes beyond the value 0.90, while none of them return this type of extreme value when they are calculated inside the WHILE WEND loop. They are all well below, in orders of magnitude that have nothing to do …
  In short, if someone has time to check what I wrote, I’m ready to answer questions 🙂 Thank you!

  // --- startperiod=20 //start period of the loop maxscale=500 //end period of the loop Step=10 //period step of the loop BandsDeviations=1.25 //bands deviation (WAVE PM) Chars= 100 //period of observation (WAVE PM) // --- psize=pointsize BandsPeriod=startperiod //first period to test is..startperiod while BandsPeriod<=maxscale do // --- wave PM indicator calculation Dev = std[BandsPeriod]*BandsDeviations Dev1 = square(Dev / psize) //temp=sqrt(average[chars](dev1))*pointsize temp=sqrt(summation[chars](dev1)/chars)*psize if temp<>0 then temp=dev/temp endif if(temp>0) then iexp=Exp(-2*temp) returnNum= (1-iexp)/(1+iexp) osc= (returnNum) else iexp=Exp(2*temp) returnNum=(iexp-1)/(1+iexp) osc= (returnNum) endif // ----- result=osc r=0 if bandsperiod=200 then r=255 endif drawtext("#result#",barindex,BandsPeriod,dialog,bold,10)coloured(r,0,0) BandsPeriod=max(startperiod,BandsPeriod+Step) //increase indicator period for next loop iteration wend return startperiod,maxscale

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

  // --- startperiod=20 //start period of the loop maxscale=500 //end period of the loop Step=10 //period step of the loop BandsDeviations=1.25 //bands deviation (WAVE PM) Chars= 100 //period of observation (WAVE PM) // ---   psize=pointsize BandsPeriod=startperiod //first period to test is..startperiod   while BandsPeriod<=maxscale do   // --- wave PM indicator calculation Dev = std[BandsPeriod]*BandsDeviations Dev1 = square(Dev / psize)   //temp=sqrt(average[chars](dev1))*pointsize temp=sqrt(summation[chars](dev1)/chars)*psize if temp<>0 then   temp=dev/temp endif   if(temp>0) then   iexp=Exp(-2*temp)   returnNum= (1-iexp)/(1+iexp)   osc= (returnNum) else   iexp=Exp(2*temp)   returnNum=(iexp-1)/(1+iexp)   osc= (returnNum) endif   // -----   result=osc   r=0 if bandsperiod=200 then   r=255 endif   drawtext("#result#",barindex,BandsPeriod,dialog,bold,10)coloured(r,0,0)   BandsPeriod=max(startperiod,BandsPeriod+Step) //increase indicator period for next loop iteration wend   return startperiod,maxscale

   

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  06/28/2019 at 9:51 AM #101592
  Nicolas

  Just made a quick test with a RSI instead of the wave PM code and the results match perfectly.. There must be something wrong in the wave PM code when it is parsed in the loop..

  // --- startperiod=20 //start period of the loop maxscale=500 //end period of the loop Step=10 //period step of the loop // --- BandsPeriod=startperiod //first period to test is..startperiod while BandsPeriod<=maxscale do // --- wave PM indicator calculation osc = rsi[bandsperiod] // ----- result=osc r=0 if bandsperiod=50 then r=255 endif drawtext("#result#",barindex,BandsPeriod,dialog,bold,10)coloured(r,0,0) BandsPeriod=max(startperiod,BandsPeriod+Step) //increase indicator period for next loop iteration wend return startperiod,maxscale

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

  // --- startperiod=20 //start period of the loop maxscale=500 //end period of the loop Step=10 //period step of the loop // ---   BandsPeriod=startperiod //first period to test is..startperiod   while BandsPeriod<=maxscale do   // --- wave PM indicator calculation osc = rsi[bandsperiod] // -----   result=osc   r=0 if bandsperiod=50 then r=255 endif   drawtext("#result#",barindex,BandsPeriod,dialog,bold,10)coloured(r,0,0)   BandsPeriod=max(startperiod,BandsPeriod+Step) //increase indicator period for next loop iteration wend   return startperiod,maxscale

   

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  06/28/2019 at 10:04 AM #101593
  Nicolas

  For people enjoying color on their charts.. RSI heatmap is working fine.
  Colors shade from red (oversold) to green (overbought) from period 20 to 500 on a 1-hour and 5-minutes charts. Not the subject of that topic, but information are valuable IMO. Trend following or mean reversion valuable infos in that heatmap.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  06/28/2019 at 11:32 AM #101607
  Nicolas

  I posted the RSI heatmap in the library, here: RSI MultiPeriods HeatMap

  Now, let’s come back to the original subject.. 😳

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  07/01/2019 at 8:33 AM #101739
  Nicolas

  To continue my monologue and for the records. The problem is located in the average of the standard deviation realized in the loop. With the AVERAGE statement or a sum of the STD divided by the period amount, the result is the same = wrong values.

  I made a report to PRT.

  1 user thanked author for this post.

  jebus89

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  07/01/2019 at 10:26 AM #101741
  jebus89

  Insane work Nicolas, thanks alot. Lets hope PRT replies within this year  then hehe

  

  jebus89

  Participant

  Topics: 118

  Replies: 586

  Been thanked: 238 times
  07/04/2019 at 5:11 PM #101983
  Vonasi

  You can make very pretty pictures if you take an average or RSI and Stochastic combined.

  //PRC_RSI and Stochastic multiperiods HeatMap | indicator //Plot an heatmap of the RSI range of periods //04.07.2019 // --- settings startperiod=2 //start period of the loop maxscale=500 //end period of the loop Step=10 //period step of the loop // --- end of settings iPeriod=startperiod //first period to test is..startperiod while iPeriod<=maxscale do // --- indicator calculation osc = (stochastic[iperiod,5] + rsi[iperiod])/2 // ----- result=osc R = max(0,50+(200-(result-50)*12)) G = max(0,50+(200+(result-50)*12)) drawtext("■",barindex,iperiod,dialog,bold,18) coloured(min(r,255),min(g,255),0) iPeriod=max(startperiod,iPeriod+Step) //increase indicator period for next loop iteration wend return startperiod,maxscale

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

  //PRC_RSI and Stochastic multiperiods HeatMap | indicator //Plot an heatmap of the RSI range of periods //04.07.2019   // --- settings startperiod=2 //start period of the loop maxscale=500 //end period of the loop Step=10 //period step of the loop // --- end of settings   iPeriod=startperiod //first period to test is..startperiod   while iPeriod<=maxscale do   // --- indicator calculation osc = (stochastic[iperiod,5] + rsi[iperiod])/2 // -----   result=osc   R = max(0,50+(200-(result-50)*12)) G = max(0,50+(200+(result-50)*12))   drawtext("■",barindex,iperiod,dialog,bold,18) coloured(min(r,255),min(g,255),0)   iPeriod=max(startperiod,iPeriod+Step) //increase indicator period for next loop iteration wend   return startperiod,maxscale

   

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1811 times
  07/04/2019 at 5:29 PM #101985
  Nicolas

  As requested, here is another pretty picture but with only a stochastic.

  Stochastic 3D heatmap - color gradient

  //PRC_CCI multiperiods HeatMap | indicator //Plot an heatmap of the CCI range of periods //04.07.2019 //Nicolas @ www.prorealcode.com //Sharing ProRealTime knowledge // --- settings startperiod=14 //start period of the loop maxscale=500 //end period of the loop Step=10 //period step of the loop // --- end of settings iPeriod=startperiod //first period to test is..startperiod while iPeriod<=maxscale do // --- indicator calculation osc = stochastic[iperiod,3](close)//cci[iperiod] // ----- //R = max(0,50-osc) //G = max(0,50+osc) R =max(0,50+(50-(osc-50)*3)) G = 100//max(0,50+(50+(osc-50)*3)) B = max(0,50+(50+(osc-50)*4)) drawtext("■",barindex,iperiod,dialog,bold,18) coloured(min(r,255),min(g,255),min(b,255)) iPeriod=max(startperiod,iPeriod+Step) //increase indicator period for next loop iteration wend return startperiod,maxscale

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

  //PRC_CCI multiperiods HeatMap | indicator //Plot an heatmap of the CCI range of periods //04.07.2019 //Nicolas @ www.prorealcode.com //Sharing ProRealTime knowledge   // --- settings startperiod=14 //start period of the loop maxscale=500 //end period of the loop Step=10 //period step of the loop // --- end of settings   iPeriod=startperiod //first period to test is..startperiod   while iPeriod<=maxscale do   // --- indicator calculation osc = stochastic[iperiod,3](close)//cci[iperiod] // -----   //R = max(0,50-osc) //G = max(0,50+osc) R =max(0,50+(50-(osc-50)*3)) G = 100//max(0,50+(50+(osc-50)*3)) B = max(0,50+(50+(osc-50)*4))   drawtext("■",barindex,iperiod,dialog,bold,18) coloured(min(r,255),min(g,255),min(b,255))   iPeriod=max(startperiod,iPeriod+Step) //increase indicator period for next loop iteration wend   return startperiod,maxscale

  We are now far from the original topic. 😳

  Anyway, I had confirmation today that the average of the ATR is the problem (the average take the last calculated data and not the one of the period calculated dynamically in the loop..). Waiting for a fix.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  07/04/2019 at 5:50 PM #101987
  Vonasi

  Glad to hear that there is a fix in the pipeline so that you can finish this indicator and add it to the library.

  Even if it turns out not to be a useful trading tool at least it will be very colourful which is what we all want in an indicator!

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1811 times
  07/04/2019 at 5:54 PM #101988
  Nicolas

  I studied the theory behind the colors and found it useful 😆

  (private joke if anyone wonder).

  1 user thanked author for this post.

  denmar

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  12/14/2019 at 11:49 AM #114860
  denmar

  Hi,

  Has this problem been resolved yet.  Sure would like to see the 3D version of this indicator.

  Thank you.

  

  denmar

  Participant

  Topics: 14

  Replies: 59

  Been thanked: 4 times
  12/16/2019 at 10:09 AM #114957
  Nicolas

  I do not think so, but I’m currently investigating in version 11..

  1 user thanked author for this post.

  denmar

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  01/01/2020 at 11:50 PM #115707
  AlgoAlex

  Has somebody eventually checked how to signal when ever period is full green or full red?

  I played around this idea but the only event I want to take out I’m unable to code it…

  

  AlgoAlex

  Participant

  Topics: 33

  Replies: 64

  Been thanked: 25 times
  01/02/2020 at 5:20 PM #115783
  Nicolas

  @denmar

  Still not unable to do it in v11, but the variables array support to come should be very helpful to make it possible.

  
  @AlexF

  Which indicator please? There are several in this topic!

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  01/02/2020 at 7:29 PM #115796
  denmar

  Thanks for update Nicholas…unfortunately I am still restricted to v 10 here in South Africa.

  

  denmar

  Participant

  Topics: 14

  Replies: 59

  Been thanked: 4 times
• Author
  Posts