Screener for my own indicator

• 08/12/2018 at 11:34 AM #78037
  malaguti

  I have developed an indicator using v10.3, very simple, but its my first one!

  what i am trying to do is now create the screener, however its not coming back with what i would expect. I want to screen new breakouts, so where the condition today exists, whereas yesterday it didn’t. although weekly, so this week vs last week

  here is the screener:

  Screener from indicator created

  myMSI2 = call MSI2 Con1 = myMSI2 = 1 and myMSI2[1] = 0.5 screener[con1]

  1 2 3 4

  myMSI2 = call MSI2 Con1 = myMSI2 = 1 and myMSI2[1] = 0.5   screener[con1]

  the indicator is here

  Indicator

  MyHigh = Highest[20](high) MyLow = Lowest[40] (low) //Ichimoku Tenkansen = (highest[9](high)+lowest[9](low))/2 Kijunsen = (highest[26](high)+lowest[26](low))/2 SpanA = (tenkansen[26]+kijunsen[26])/2 SpanB = (highest[52](high[26])+lowest[52](low[26]))/2 //MA Settings MyMA = WilderAverage[30](close) Long1 = High>MyHigh[1] Long2 = close > SpanA and close > SpanB Long3 = Close>Close[1] and close>MyHigh[1] Long4 = Close>MyMA Stage2 = Long1 and Long2 and Long3 and Long4 Stage3 = Close<MyMA and Direction[1]=1 Short1 = Close<MyLow[1] Short2 = close < SpanA and close < SpanB Short3 = Close<Close[1] and close<MyLow[1] Stage4 = Short1 and short2 and short3 Stage1 = Close>MyMA and Direction[1]=-1 if stage2 then direction=1 elsif stage3 then direction=0.5 elsif stage4 then direction=-1 elsif stage1 then direction=-0.5 endif return direction as "MSI"

  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

  MyHigh = Highest[20](high) MyLow = Lowest[40] (low) //Ichimoku Tenkansen = (highest[9](high)+lowest[9](low))/2 Kijunsen = (highest[26](high)+lowest[26](low))/2 SpanA = (tenkansen[26]+kijunsen[26])/2 SpanB = (highest[52](high[26])+lowest[52](low[26]))/2 //MA Settings MyMA = WilderAverage[30](close)     Long1 = High>MyHigh[1] Long2 = close > SpanA and close > SpanB Long3 = Close>Close[1] and close>MyHigh[1] Long4 = Close>MyMA Stage2 = Long1 and Long2 and Long3 and Long4   Stage3 = Close<MyMA and Direction[1]=1   Short1 = Close<MyLow[1] Short2 = close < SpanA and close < SpanB Short3 = Close<Close[1] and close<MyLow[1] Stage4 = Short1 and short2 and short3   Stage1 = Close>MyMA and Direction[1]=-1   if stage2 then direction=1 elsif stage3 then direction=0.5 elsif stage4 then direction=-1 elsif stage1 then direction=-0.5 endif   return direction as "MSI"

  the indicator when plotted has a 1,0.5, -0.5 or -1 value. Im just trying to pick up the 1 where last week it wasn’t however im getting back items where there is no 1 but still 0.5

   

   

   

   

  

  could someone help where im going wrong?

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
  08/14/2018 at 11:07 AM #78134
  malaguti

  If I run the screener on the predefined UK stocks list, using the weekly selection criteria my results are attached:

  none of them fit the criteria of the screener.

  Does ProRealTime just not like the weekly criteria?

   

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
  08/14/2018 at 11:55 AM #78136
  robertogozzi

  It works, but only few results are returned.

  Try to san ALL instruments to see how many results you get, in different TFs.

  

  robertogozzi

  Moderator

  Topics: 69

  Replies: 17131

  Been thanked: 4030 times
  08/14/2018 at 12:17 PM #78138
  robertogozzi

  I suggest that you use your indicator as a strategy and debug it with ProOrder, changing GRAPHed variables to watch all of them:

  MyHigh = Highest[20](high) MyLow = Lowest[40](low) //Ichimoku Tenkansen = (highest[9](high)+lowest[9](low))/2 Kijunsen = (highest[26](high)+lowest[26](low))/2 SpanA = (tenkansen[26]+kijunsen[26])/2 SpanB = (highest[52](high[26])+lowest[52](low[26]))/2 //MA Settings MyMA = WilderAverage[30](close) Long1 = High>MyHigh[1] Long2 = close > SpanA and close > SpanB Long3 = Close>Close[1] and close>MyHigh[1] Long4 = Close>MyMA Stage2 = Long1 and Long2 and Long3 and Long4 Stage3 = Close<MyMA and Direction[1]=1 Short1 = Close<MyLow[1] Short2 = close < SpanA and close < SpanB Short3 = Close<Close[1] and close<MyLow[1] Stage4 = Short1 and short2 and short3 Stage1 = Close>MyMA and Direction[1]=-1 if stage2 then direction=1 elsif stage3 then direction=0.5 elsif stage4 then direction=-1 elsif stage1 then direction=-0.5 endif graph MyHigh graph MyLow graph MyMA graph close graph high buy at -close limit

  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

  MyHigh    = Highest[20](high) MyLow     = Lowest[40](low) //Ichimoku Tenkansen = (highest[9](high)+lowest[9](low))/2 Kijunsen  = (highest[26](high)+lowest[26](low))/2 SpanA     = (tenkansen[26]+kijunsen[26])/2 SpanB     = (highest[52](high[26])+lowest[52](low[26]))/2   //MA Settings MyMA      = WilderAverage[30](close) Long1     = High>MyHigh[1] Long2     = close > SpanA and close > SpanB Long3     = Close>Close[1] and close>MyHigh[1] Long4     = Close>MyMA   Stage2    = Long1 and Long2 and Long3 and Long4 Stage3    = Close<MyMA and Direction[1]=1 Short1    = Close<MyLow[1] Short2    = close < SpanA and close < SpanB Short3    = Close<Close[1] and close<MyLow[1] Stage4    = Short1 and short2 and short3 Stage1    = Close>MyMA and Direction[1]=-1   if stage2 then direction=1 elsif stage3 then direction=0.5 elsif stage4 then direction=-1 elsif stage1 then direction=-0.5 endif   graph MyHigh graph MyLow graph MyMA graph close graph high   buy at -close limit

  the last line is a dummy one required by ProOrder.

   

  

  robertogozzi

  Moderator

  Topics: 69

  Replies: 17131

  Been thanked: 4030 times
  08/14/2018 at 12:33 PM #78140
  malaguti

  robertogozzi wrote:

  I suggest that you use your indicator as a strategy and debug it with ProOrder, changing GRAPHed variables to watch all of them:

  MyHigh = Highest[20](high)<br /> MyLow = Lowest[40](low)<br /> //Ichimoku<br /> Tenkansen = (highest[9](high)+lowest[9](low))/2<br /> Kijunsen = (highest[26](high)+lowest[26](low))/2<br /> SpanA = (tenkansen[26]+kijunsen[26])/2<br /> SpanB = (highest[52](high[26])+lowest[52](low[26]))/2</p> <p>//MA Settings<br /> MyMA = WilderAverage[30](close)</p> <p>Long1 = High>MyHigh[1]<br /> Long2 = close > SpanA and close > SpanB<br /> Long3 = Close>Close[1] and close>MyHigh[1]<br /> Long4 = Close>MyMA</p> <p>Stage2 = Long1 and Long2 and Long3 and Long4</p> <p>Stage3 = Close<MyMA and Direction[1]=1</p> <p>Short1 = Close<MyLow[1]<br /> Short2 = close < SpanA and close < SpanB<br /> Short3 = Close<Close[1] and close<MyLow[1]<br /> Stage4 = Short1 and short2 and short3</p> <p>Stage1 = Close>MyMA and Direction[1]=-1</p> <p>if stage2 then<br /> direction=1<br /> elsif stage3 then<br /> direction=0.5<br /> elsif stage4 then<br /> direction=-1<br /> elsif stage1 then<br /> direction=-0.5<br /> endif</p> <p>graph MyHigh<br /> graph MyLow<br /> graph MyMA<br /> graph close<br /> graph high</p> <p>buy at -close limit

  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

  MyHigh = Highest[20](high) MyLow = Lowest[40](low) //Ichimoku Tenkansen = (highest[9](high)+lowest[9](low))/2 Kijunsen = (highest[26](high)+lowest[26](low))/2 SpanA = (tenkansen[26]+kijunsen[26])/2 SpanB = (highest[52](high[26])+lowest[52](low[26]))/2 //MA Settings MyMA = WilderAverage[30](close) Long1 = High>MyHigh[1] Long2 = close > SpanA and close > SpanB Long3 = Close>Close[1] and close>MyHigh[1] Long4 = Close>MyMA Stage2 = Long1 and Long2 and Long3 and Long4 Stage3 = Close<MyMA and Direction[1]=1 Short1 = Close<MyLow[1] Short2 = close < SpanA and close < SpanB Short3 = Close<Close[1] and close<MyLow[1] Stage4 = Short1 and short2 and short3 Stage1 = Close>MyMA and Direction[1]=–1 if stage2 then direction=1 elsif stage3 then direction=0.5 elsif stage4 then direction=–1 elsif stage1 then direction=–0.5 endif graph MyHigh graph MyLow graph MyMA graph close graph high buy at –close limit

  the last line is a dummy one required by ProOrder.

  Its not so much the volume, its just they are all inaccurate. I will try your suggestion however on proorder, I haven’t tried that yet, hoping it was an obvious error in my indicator

   

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
  08/14/2018 at 2:00 PM #78146
  malaguti

  So I created the probacktest, and graphed the indicator, and it buys and sells exactly where it should

  the screener just cant pick up the right information however

  I am keeping it nice and simple to just return where the condition is true, in this case where the indicator value is -0.5

  it returns a stock where its -1

  probacktest works perfectly, just the screener cant work it out

  this time I also included everything within the screener rather than “calling” it and still incorrect results

  any ideas?

   

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
  08/14/2018 at 2:49 PM #78151
  robertogozzi

  As an indicator it returns all values and they are pletted correctly under the chart.

  Using the same code as a screener, simply replacing return with screener:

  screener [direction] (direction as "Dir")

  1	screener [direction] (direction as "Dir")

  only returns 1 and 0.5!

  I launched it on a 1-minute TF, to watch any fast update, but I did not see other that 1’s and 0.5’s.

   

  

  robertogozzi

  Moderator

  Topics: 69

  Replies: 17131

  Been thanked: 4030 times
  08/14/2018 at 3:14 PM #78152
  malaguti

  Yes, and when you request to just bring back all with a 1, it brings back -0.5!

  I want to look for changes, just as probuilder was able to handle, when direction =1 and direction[1]=0.5

  

   

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
  08/14/2018 at 5:28 PM #78166
  malaguti

  here the screener thinks I’m on a 1, whereas the indicator shows the correct value, which is -0.5

  in fact they all show as 1

  

  any help would be appreciated

   

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
  08/15/2018 at 6:13 PM #78304
  malaguti

  OK, decided to try again and perhaps narrow down where the screener is going wrong and wanted to just test any standard code in case it was my own poor coding being such a novice.

  the below screener looks at 4 conditions, and it still brings back incorrect results from standard code

  standard code

  MyHigh = Highest[20](high) //MyLow = Lowest[40] (low) //Ichimoku Tenkansen = (highest[9](high)+lowest[9](low))/2 Kijunsen = (highest[26](high)+lowest[26](low))/2 SpanA = (tenkansen[26]+kijunsen[26])/2 SpanB = (highest[52](high[26])+lowest[52](low[26]))/2 //MA Settings MyMA = WilderAverage[30](close) Long1 = High>MyHigh[1] Long2 = close > SpanA and close > SpanB Long3 = Close>Close[1] and close>MyHigh[1] Long4 = Close>MyMA Stage2 = Long1 and Long2 and Long3 and Long4 SCREENER [Stage2] (Stage2 AS "Stage2")

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

  MyHigh = Highest[20](high) //MyLow = Lowest[40] (low) //Ichimoku Tenkansen = (highest[9](high)+lowest[9](low))/2 Kijunsen = (highest[26](high)+lowest[26](low))/2 SpanA = (tenkansen[26]+kijunsen[26])/2 SpanB = (highest[52](high[26])+lowest[52](low[26]))/2 //MA Settings MyMA = WilderAverage[30](close)     Long1 = High>MyHigh[1] Long2 = close > SpanA and close > SpanB Long3 = Close>Close[1] and close>MyHigh[1] Long4 = Close>MyMA Stage2 = Long1 and Long2 and Long3 and Long4   SCREENER [Stage2] (Stage2 AS "Stage2")

  PRTs screener is not able to bring back the correct results even from standard code

  in the attached chart, the high is clearly not greater than last week’s high, so Long 1 isn’t satisfied and Long3 also isn’t satisfied

  I’ve overlaid the donchian channel, highest high for 20 periods. it brings back Vedanta as saying its higher than last period, whereas it clearly isnt

  so its not certainly not my indicator that is the issue. Nobody else has similar results?

   

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
  08/15/2018 at 6:24 PM #78310
  malaguti

  lets make it real simple just one condition..

  just one condition

  MyHigh = Highest[20](high) Long1 = High>MyHigh[1] SCREENER [Long1] (Long1 AS "Higher")

  1 2 3 4

  MyHigh = Highest[20](high)   Long1 = High>MyHigh[1] SCREENER [Long1] (Long1 AS "Higher")

  on the weekly screener for predefined UK200 stocks..brings back some very odd results. definitely not higher!

   

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
  08/16/2018 at 11:35 AM #78356
  robertogozzi

  Converting you screener to an indicator colouring the background when the result is true

  MyHigh = Highest[20](high) Long1 = High>MyHigh[1] IF Long1 THEN backgroundcolor(0,255,0) endif return

  1 2 3 4 5 6

  MyHigh = Highest[20](high) Long1 = High>MyHigh[1] IF Long1 THEN    backgroundcolor(0,255,0) endif return

  then launching your scan on UK stocks (weekly TF), results can be seen from attached pics, the first one is correct while the second one isn’t.

  The only explanation is that you are using end-of-day data, because the screener, as well as the indicator, are both correct if you don’t look at the current bar!

   

  

  robertogozzi

  Moderator

  Topics: 69

  Replies: 17131

  Been thanked: 4030 times
  08/16/2018 at 11:41 AM #78359
  robertogozzi

  Well… I’m afraid there’s something wrong elsewhere, because that does not explain why in my first screenshot, on July 16th or 23rd…, the background is coloured while there’e no NEW high!

  

  robertogozzi

  Moderator

  Topics: 69

  Replies: 17131

  Been thanked: 4030 times
  08/16/2018 at 12:35 PM #78362
  malaguti

  Thanks Roberto, yes I’ve just marked up where it shouldn’t have occurred. It is odd. and I do just use just end of day, as I’m only interested in weekly timeframes

  I’ve raised a support ticket but I do appreciate your responses, and at least confirms I’m not losing my mind!

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
  08/16/2018 at 2:05 PM #78368
  malaguti

  The plot thickens…

  Long = low<low[1] if long then backgroundcolor (0,255,0) endif return long

  1 2 3 4 5

  Long = low<low[1] if long then backgroundcolor (0,255,0) endif return long

  wanting to check now very very simple indicator checking one high to another or one low to another, I’m finding that PRT isn’t updating the indicator. ie if I switch from one timescale to another I get incorrect results, unless I go into modify an indicator and then choose to re-apply the same indicator. its almost as though when switching that PRT isn’t catching up when switching and reading the code

  attachment 1 is supposed to highlight just where the low is lower than last period, compared to the exact same indicator after ive modified the indicator and added the indicator to the chart again.

  its the exact same timeframe, same chart, two different results after reapplying

  ive tried closing down the indicator, after creating it. but I wonder is this why the screener is coming back with the odd results..

   

  

  malaguti

  Participant

  Topics: 2

  Replies: 16

  Been thanked: 0 times
• Author
  Posts