WilderAverage Calculation

• 11/27/2016 at 7:12 PM #17320
  Fabian

  Hello,

  i want to rebuild the RSI Indikator of ProRealtime in Java. The ProRealtime Example is very helpful:

  REM Computes the daily variations UP = MAX(0, close - close[1]) DOWN = MAX(0, close[1] - close) REM Computes the moving average of gains on positive days REM and losses on negative days upMA = wilderAverage[p](UP) downMA = wilderAverage[p](DOWN) REM Now we can compute the RS RS = upMA / downMA REM And finally the RSI myRSI = 100 - 100 / (1 + RS) RETURN myRSI AS "Relative Strength Index"

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

  REM Computes the daily variations   UP = MAX(0, close - close[1]) DOWN = MAX(0, close[1] - close)   REM Computes the moving average of gains on positive days REM and losses on negative days   upMA = wilderAverage[p](UP) downMA = wilderAverage[p](DOWN)   REM Now we can compute the RS   RS = upMA / downMA   REM And finally the RSI   myRSI = 100 - 100 / (1 + RS)   RETURN myRSI AS "Relative Strength Index"

  But i dont know how to calculate the wilderAverage. What code/formula belongs to the wilderAverage?

   

  Thank you for help!

  

  Fabian

  Participant

  Topics: 3

  Replies: 4

  Been thanked: 0 times
  11/27/2016 at 7:30 PM #17321
  Nicolas

  Found this formula on Internet :

  EMA formula = price today * K + EMA yesterday * (1-K) where K = 2 / (N+1)

  Wilder EMA formula = price today * K + EMA yesterday (1-K) where K =1/N

  Where N = the number of periods.

  Hope it helps you.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  11/28/2016 at 10:14 PM #17421
  Fabian

  Not really 🙁

  Easy Example with 3 Candles:

  Candle1: Close 4897
  Candle2: Close 4934
  Candle3: Close 4951

  Value for Candle3

  UP = MAX(0, close – close[1]) –> 4951 – 4934 –> 17
  DOWN = MAX(0, close[1] – close) –> 4934 – 4951 –> -17

  upMA = wilderAverage[p](UP) –> wilderAverage[1](17) –> 17
  downMA = wilderAverage[p](DOWN) –> wilderAverage[1](-17) –> 0

  upMA = wilderAverage[p](UP) –> wilderAverage[2](17) –> 20,87160 ???
  downMA = wilderAverage[p](DOWN) –> wilderAverage[2](-17) –> 0,20206 ???

  upMA = wilderAverage[p](UP) –> wilderAverage[3](17) –> 19,65567 ???
  downMA = wilderAverage[p](DOWN) –> wilderAverage[3](-17) –> 0,71841 ???

  I don’t get it…

  The difference of the last two candle’s I have to put in the wilderAverage?

  Maybe I can get the Code from ProRealtime as the RSI Example?

  

  Fabian

  Participant

  Topics: 3

  Replies: 4

  Been thanked: 0 times
  11/28/2016 at 10:19 PM #17424
  Nicolas

  I’m curious, why do you want to rebuild it in Java?

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  11/28/2016 at 11:03 PM #17430
  Fabian

  I’m a customer of IG Markets. I trade with the java api from IG Markets. But i get the signals manually from ProRealtime (with IG Data). The IG api has the possibility to get candle Stick data. Therefore i want to automate the generation of signals.

  Thats the reason why i need the correct calculation of RSI from ProRealtime in my java application. Specially the calculation of the wilderaverage…the rest calculation of RSI is easy and clear.

  I hope you understand…if not i have to post it in German.

  

  Fabian

  Participant

  Topics: 3

  Replies: 4

  Been thanked: 0 times
  11/29/2016 at 9:11 AM #17447
  Nicolas

  That’s ok, I finally found the good formula and successfully recoded the same formula as the PRT one, here is the code:

  //WilderMA is calculated for periods "n" as follows: // //Wilder MA = ( Previous Wilder MA * ( n - 1 ) + DataSeries Value ) / n // // //where, // //n = number of periods //DataSeries Value = data you wish to average once WEMA = close price = close N = 20 if barindex>N THEN WEMA = (WEMA[1]*(N-1)+price)/N endif RETURN WEMA

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

  //WilderMA is calculated for periods "n" as follows: // //Wilder MA = ( Previous Wilder MA * ( n - 1 ) + DataSeries Value ) / n // // //where, // //n = number of periods //DataSeries Value = data you wish to average   once WEMA = close price = close N = 20   if barindex>N THEN WEMA = (WEMA[1]*(N-1)+price)/N endif   RETURN WEMA

   

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  11/29/2016 at 6:25 PM #17498
  Fabian

  I dont get it 🙁

  here is my code to check the wilderAverage (its form the RSI example)

  REM Computes the daily variations UP = MAX(0, close - close[1]) DOWN = MAX(0, close[1] - close) REM Computes the moving average of gains on positive days REM and losses on negative days upMA = wilderAverage[p](UP) downMA = wilderAverage[p](DOWN) RETURN upMA,downMA

  1 2 3 4 5 6 7 8 9 10 11 12

  REM Computes the daily variations   UP = MAX(0, close - close[1]) DOWN = MAX(0, close[1] - close)   REM Computes the moving average of gains on positive days REM and losses on negative days   upMA = wilderAverage[p](UP) downMA = wilderAverage[p](DOWN)   RETURN upMA,downMA

  Your code returns different values. I think the wilderAverage of ProRealtime returns the differents between the close prices, but i dont know how they are calculate…

  Thanks for help!

  

  Fabian

  Participant

  Topics: 3

  Replies: 4

  Been thanked: 0 times
  11/29/2016 at 8:14 PM #17514
  Nicolas

  The code I gave you return exactly the same values as the one of the platform, did you try it at least?

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  11/29/2016 at 10:02 PM #17518
  Fabian

  Yes i tried…

  here is my result.

  

  Fabian

  Participant

  Topics: 3

  Replies: 4

  Been thanked: 0 times
  03/15/2021 at 10:49 AM #164205
  Daniele Maddaluno

  I know it’s a really old post.
  Anyway I’ll add the RSI with wilderAverage recoded here for people who will need that:

  RSI with wilderaverage recoded inside

  up = max(0, close - close[1]) dn = max(0, close[1] - close) once mmUp = 0 once mmDn = 0 if barindex>1 then // wilderAverage formulas mmUp = mmUp[1] + (up - mmUp[1])/p mmDn = mmDn[1] + (dn - mmDn[1])/p endif rs = mmUp / mmDn rsindex = 100 - 100 / (1 + rs) return rsindex as "Relative Strength Index"

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

  up = max(0, close - close[1]) dn = max(0, close[1] - close)   once mmUp = 0 once mmDn = 0   if barindex>1 then    // wilderAverage formulas    mmUp = mmUp[1] + (up - mmUp[1])/p    mmDn = mmDn[1] + (dn - mmDn[1])/p endif   rs = mmUp / mmDn rsindex = 100 - 100 / (1 + rs)   return rsindex as "Relative Strength Index"

  It needs some bars to converge to the PRT RSI indicator.

  1 user thanked author for this post.

  Nicolas

  

  Daniele Maddaluno

  Participant

  Topics: 8

  Replies: 67

  Been thanked: 14 times
• Author
  Posts