Simple Hull MA screener won't work

• 06/17/2019 at 3:34 PM #100889
  JMat45

  Hi there, I’ve had issues with screeners and using the ‘Average’ command before (apparently ProScreener doesn’t like the second argument in the parenthesis so much…), but thought they would have been fixed by now. Can anyone spot the problem with this (doesn’t return any results on any list of instruments)?

  Hull MA turn-up

  HullP1 = 6 RSIper = 6 MyRSI = RSI[RSIper](Close) HullMA1 = Average[HullP1, 8](Close) IF (HullMA1 > HullMA1[1]) AND (HullMA1[1] < HullMA1[2]) THEN HullMA1turnup = 1 ELSE HullMA1turnup = 0 ENDIF Hullturnup = HullMA1turnup OR HullMA1turnup[1] OR HullMA1turnup[2] OR HullMA1turnup[3] SCREENER[Hullturnup](MyRSI)

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

  HullP1 = 6 RSIper = 6   MyRSI = RSI[RSIper](Close)   HullMA1 = Average[HullP1, 8](Close)   IF (HullMA1 > HullMA1[1]) AND (HullMA1[1] < HullMA1[2]) THEN HullMA1turnup = 1 ELSE HullMA1turnup = 0 ENDIF   Hullturnup = HullMA1turnup OR HullMA1turnup[1] OR HullMA1turnup[2] OR HullMA1turnup[3]   SCREENER[Hullturnup](MyRSI)

  It seems ProScreener will only return results up to the value of 6 as the MA type…

  Even when I manually calculate the Hull MA I don’t get any results, even though translating to an indicator to obtain e.g. a histogram when the condition is true works just fine:

  Manual Hull MA turn-up code

  HullP1 = 6 RSIper = 6 MyRSI = RSI[RSIper](Close) a1 = (HullP1/2)-((HullP1 MOD 2)/2) b1 = SQRT(HullP1) MOD HullP1 c1 = WeightedAverage[a1](Close) d1 = WeightedAverage[HullP1](Close) HullMA1 = WeightedAverage[b1]((2*c1)-d1) IF (HullMA1 > HullMA1[1]) AND (HullMA1[1] < HullMA1[2]) THEN HullMA1turnup = 1 ELSE HullMA1turnup = 0 ENDIF SCREENER[HullMA1turnup](MyRSI)

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

  HullP1 = 6 RSIper = 6   MyRSI = RSI[RSIper](Close)   a1 = (HullP1/2)-((HullP1 MOD 2)/2) b1 = SQRT(HullP1) MOD HullP1   c1 = WeightedAverage[a1](Close) d1 = WeightedAverage[HullP1](Close)   HullMA1 = WeightedAverage[b1]((2*c1)-d1)   IF (HullMA1 > HullMA1[1]) AND (HullMA1[1] < HullMA1[2]) THEN HullMA1turnup = 1 ELSE HullMA1turnup = 0 ENDIF   SCREENER[HullMA1turnup](MyRSI)

  Doesn’t fill me with confidence in ProRealTime as a platform if simple things like this can’t be programmed. I know v11 is coming out soon with a designated HullAverage instruction, but still…

  Thanks in advance for your help.

   

  

  JMat45

  Participant

  Topics: 6

  Replies: 23

  Been thanked: 2 times
  06/17/2019 at 9:45 PM #100923
  GraHal

  I get ProOrder / Auto-Systems working okay for MA[20,7] and MA[20,8] but it seems MA type 7 and 8 are not supposed to work in PRT v 10.3.

  Maybe it is just as a Screener that MA Type 7 and 8 don’t work in V 10.3

  MA type 7 and 8 do work in PRT v 11 … so I am told, as I don’t have v 11.

  

  GraHal

  Participant

  Topics: 211

  Replies: 8783

  Been thanked: 2183 times
  06/17/2019 at 10:20 PM #100927
  robertogozzi

  Install yjis indicator on your chart, to find out if there have been signals before:

  HullP1 = 6 //RSIper = 6 // //MyRSI = RSI[RSIper](Close) a1 = (HullP1/2)-((HullP1 MOD 2)/2) b1 = SQRT(HullP1) MOD HullP1 c1 = WeightedAverage[a1](Close) d1 = WeightedAverage[HullP1](Close) HullMA1 = WeightedAverage[b1]((2*c1)-d1) IF (HullMA1 > HullMA1[1]) AND (HullMA1[1] < HullMA1[2]) THEN HullMA1turnup = 1 ELSE HullMA1turnup = 0 ENDIF RETURN HullMA1turnup

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

  HullP1 = 6 //RSIper = 6 // //MyRSI = RSI[RSIper](Close) a1 = (HullP1/2)-((HullP1 MOD 2)/2) b1 = SQRT(HullP1) MOD HullP1 c1 = WeightedAverage[a1](Close) d1 = WeightedAverage[HullP1](Close) HullMA1 = WeightedAverage[b1]((2*c1)-d1) IF (HullMA1 > HullMA1[1]) AND (HullMA1[1] < HullMA1[2]) THEN    HullMA1turnup = 1 ELSE    HullMA1turnup = 0 ENDIF RETURN HullMA1turnup

   

  

  robertogozzi

  Moderator

  Topics: 69

  Replies: 17131

  Been thanked: 4030 times
  06/17/2019 at 10:52 PM #100930
  JMat45

  Yes I did that thanks, Roberto, as alluded to in my post (“even though translating to an indicator to obtain e.g. a histogram when the condition is true works just fine”). I can confirm it works fine as an indicator, yet ProScreener doesn’t pick it up…

  The screener is designed for cryptos so am not so interested in backtesting on ProOrder, GraHal, but thanks for your observation.

  

  JMat45

  Participant

  Topics: 6

  Replies: 23

  Been thanked: 2 times
  06/17/2019 at 11:03 PM #100931
  robertogozzi

  With HULL periods = 100 it works.

  I also added line 19 and modfied line 20 to make a signal valid for longer than just one bar:

  LookBack = 1 //make it valid for how many bars you may need SCREENER[summation[LookBack](HullMA1turnup)](MyRSI)

  1 2	LookBack = 1       //make it valid for how many bars you may need SCREENER[summation[LookBack](HullMA1turnup)](MyRSI)

   

  

  robertogozzi

  Moderator

  Topics: 69

  Replies: 17131

  Been thanked: 4030 times
  06/17/2019 at 11:55 PM #100934
  JMat45

  Thanks for that, Roberto. A 100-period lookback is not ideal – it should work for all (positive/non-zero) values.

  I have tried the Summation command before to detect signals from previous bars, but for the code I posted above that did not work either.

  

  JMat45

  Participant

  Topics: 6

  Replies: 23

  Been thanked: 2 times
  06/18/2019 at 10:09 AM #100945
  GraHal

  JMat45 wrote:

  A 100-period lookback is not ideal

  What myself and Roberto were trying to provide is information to show if Hull MA works in a ProScreener as your original post suspected that Hull MA didnt work in ProScreener.

  Roberto’s code / results proves that HullMA does work in ProScreener.

  

  GraHal

  Participant

  Topics: 211

  Replies: 8783

  Been thanked: 2183 times
  06/18/2019 at 10:24 AM #100946
  JMat45

  Yes, but only with certain lookback periods (beyond what is a high initial value). I think most people would regard that as sub-optimal.

  

  JMat45

  Participant

  Topics: 6

  Replies: 23

  Been thanked: 2 times
  06/18/2019 at 10:55 AM #100948
  GraHal

  JMat45 wrote:

  I think most people would regard that as sub-optimal.

  Maybe and we take / agree your point,  but Roberto code shows that ProScreener can read the Hull MA type 8 … which your original post read / doubted that HullMA could be used at all at any lookback period / in any way at all in ProScreener?.

  I guess now we are onto the next phase of coding … to get a Screener that gives you want you want (now we know / have proven that Hull MA works in ProScreener).

  

  GraHal

  Participant

  Topics: 211

  Replies: 8783

  Been thanked: 2183 times
  06/18/2019 at 10:58 AM #100949
  JMat45

  Thank you, GraHal.

  

  JMat45

  Participant

  Topics: 6

  Replies: 23

  Been thanked: 2 times
• Author
  Posts