Screener Help for Nicolas's Standard Deviation/Error LR Channel

• 02/09/2020 at 1:09 PM #119085
  Newbie

  Greetings team. I’ve been working on back testing with Nicolas’s Indicator here: https://www.prorealcode.com/prorealtime-indicators/standard-deviation-standard-error-linear-regression-channel/

  I’d like to develop a screener along with another indicator in order to share with you all, however, I’m a little stuck with the conditions. Ideally this is a question directly to Nicolas. But anyone who can help, please feel free. I can do the rest of the work on my own I think.

  As a visual indicator, this is amazing. But how does one detect digitally, when price (close) is above (>) the top line (+1) of the channel? I’m struggling with converting the drawn-line to a number or fixed point to use for a reference.

  Ideally I want to make the top line of the channel, indicator 1 and the bottom line, indicator 2 respectively.

  I think this has real potential…

   

   

  

  Newbie

  Participant

  Topics: 4

  Replies: 11

  Been thanked: 1 time
  02/09/2020 at 1:31 PM #119086
  Vonasi

  I’m guessing that you want to know if the current bar is above or below the lines and not if a candle in the lookback period has been above or below  the lines?

  If so then a+b*0+dat is the upper line y value and a+b*0–dat is the lower line y value for the current bar.

  1 user thanked author for this post.

  Newbie

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1811 times
  02/09/2020 at 2:57 PM #119092
  Newbie

  absolutely that… Thanks, I’ll have a play round with that now and see the results. Thank you, Craig 🙂

  

  Newbie

  Participant

  Topics: 4

  Replies: 11

  Been thanked: 1 time
  02/09/2020 at 5:48 PM #119101
  Nicolas

  I’m sure that I made that screener already, let me check in the forums.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  02/09/2020 at 5:55 PM #119103
  Nicolas

  I can’t find it now.. anyway, what you are looking for is if the price is exceding a linear regression + standard deviation factorized, should be coded like this:

  test = high > linearregression[lookback]+std[lookback]*NbDeviation OR low < linearregression[lookback]-std[lookback]*NbDeviation screener[test]

  1 2 3

  test = high > linearregression[lookback]+std[lookback]*NbDeviation OR low < linearregression[lookback]-std[lookback]*NbDeviation   screener[test]

  (not tested)

  1 user thanked author for this post.

  Newbie

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  02/10/2020 at 5:33 PM #119208
  Newbie

  Okay guys… Thank you so far. I’ve managed to learn a bit more from what you’ve both told me and I’ve merged both comments into one screener prototype. However… the MAX units I can set for the lookback is 254, ideally I’d like this to be 300.

  I’m enclosing the code below for assistance. (Please keep in mind I’m a real newcomer, but I really am interested in learning more).

  Screener Prototype - Nicolas's PRC LR STD/STE channel

  // --- settings lookback= 254 //channel period ChannelType = 1 //1= Standard Deviation ; 2= Standard Erro NbDeviation = 1 //Deviation multiplier // --- end of settings sumx = 0 sumy = 0 sumxy = 0 sumx2 = 0 for cmpt = lookback downto 0 do tmpx = cmpt tmpy = close[cmpt] sumy = sumy+tmpy sumx = sumx+tmpx sumx2 = sumx2 + (tmpx*tmpx) sumxy = sumxy + (tmpy*tmpx) next n = lookback+1 if (sumx2 = sumx * sumx) then // protection to avoid infinite values b = sumxy - sumx * sumy else b = (n * sumxy - sumx * sumy) / (n * sumx2 - sumx * sumx) endif a = (sumy - b * sumx) / n //channel if ChannelType = 1 then //Standard Deviation dat = std[lookback]*NbDeviation else dat = ste[lookback]*NbDeviation endif y = a+b*0+dat z = a+b*0-dat test = close > y OR close < z screener[test]

  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

  // --- settings lookback= 254 //channel period ChannelType = 1 //1= Standard Deviation ; 2= Standard Erro NbDeviation = 1 //Deviation multiplier // --- end of settings   sumx = 0 sumy = 0 sumxy = 0 sumx2 = 0   for cmpt = lookback downto 0 do tmpx = cmpt tmpy = close[cmpt] sumy = sumy+tmpy sumx = sumx+tmpx sumx2 = sumx2 + (tmpx*tmpx) sumxy = sumxy + (tmpy*tmpx) next   n = lookback+1   if (sumx2 = sumx * sumx) then // protection to avoid infinite values b = sumxy - sumx * sumy else b = (n * sumxy - sumx * sumy) / (n * sumx2 - sumx * sumx) endif a = (sumy - b * sumx) / n   //channel if ChannelType = 1 then //Standard Deviation dat = std[lookback]*NbDeviation else dat = ste[lookback]*NbDeviation endif   y = a+b*0+dat z = a+b*0-dat   test = close > y OR close < z screener[test]

  Thanks for your help and patience, Craig 🙂

  

  Newbie

  Participant

  Topics: 4

  Replies: 11

  Been thanked: 1 time
  02/10/2020 at 7:27 PM #119231
  Vonasi

  Yes 254 is the maximum lookback allowed in screeners unfortunately.

  Perhaps Nicolas’s simpler version returns the same results? Otherwise that is a screener based on that indicator.

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1811 times
  02/11/2020 at 10:17 AM #119304
  Nicolas

  Indeed, because in realtime we don’t have to compute the affine function in order to plot a line in the past, the last point of the channel is enough.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  03/02/2020 at 8:01 PM #121010
  Newbie

  Hi Nicolas

  I know you’re constantly bombarded with questions, however, I was wondering if there’s a way to code that both values must be “above” the top line. (Even though I know the line is only a drawn on object)! I’m working on this as developing the indicator for multi-timeframe, but then also for use later in a strategy to share. (I just want it to be better before sharing.

  So far I have…

  Multi-TF Channel Detection Screener

  Timeframe (1 hour) test1 = high > linearregression[254]+std[254]*1 OR low < linearregression[254]-std[254]*1 Timeframe (15 Minute) test2 = high > linearregression[254]+std[254]*1 OR low < linearregression[254]-std[254]*1 screener[test1 AND test2]

  1 2 3 4 5 6 7 8

  Timeframe (1 hour)   test1 = high > linearregression[254]+std[254]*1 OR low < linearregression[254]-std[254]*1   Timeframe (15 Minute) test2 = high > linearregression[254]+std[254]*1 OR low < linearregression[254]-std[254]*1 screener[test1 AND test2]

  However, currently this would show when say the one hour is Above the Top line but the default or 15m is below the bottom line. I know that this is correct. But it’s not the outcome I wish to return.

  

  Newbie

  Participant

  Topics: 4

  Replies: 11

  Been thanked: 1 time
  03/02/2020 at 8:06 PM #121011
  Newbie

  I may have just answered my own question on this… If it’s not this, then any help would be received with gratitude.

  MTF Channel modification

  Timeframe (1 hour) test1 = high > linearregression[254]+std[254]*1 test2 = low < linearregression[254]-std[254]*1 Timeframe (15 Minute) test3 = high > linearregression[254]+std[254]*1 test4 = low < linearregression[254]-std[254]*1 screener [(test1 AND test2) OR (test3 AND test4)]

  1 2 3 4 5 6 7 8 9 10 11

  Timeframe (1 hour)   test1 = high > linearregression[254]+std[254]*1   test2 = low < linearregression[254]-std[254]*1   Timeframe (15 Minute) test3 = high > linearregression[254]+std[254]*1 test4 = low < linearregression[254]-std[254]*1 screener [(test1 AND test2) OR (test3 AND test4)]

   

  

  Newbie

  Participant

  Topics: 4

  Replies: 11

  Been thanked: 1 time
  03/03/2020 at 9:36 AM #121033
  Nicolas

  You have indeed answered your own question 🙂 That last screener code seems correct to me!

  1 user thanked author for this post.

  Newbie

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
• Author
  Posts