Weight averaged ADXR strategy

• 11/06/2017 at 11:14 AM #51604
  gabri

  This strategy is based on the modified version of the Directional Movement indicator I posted here. The strategy has to be backtested but it always gives above 70% of winning trades with a decent gain/loss ratio. The average gain per trade can be improved by changing the stop loss and the ADXR threshold.

  Blue skies.

  //tipomm=2 //x=5 - I normally use a number between 5 and 10 DEFPARAM CumulateOrders = False // Entering long ignored, ignored, mioDI, ignored, mioADXR, ignored = CALL "PRT - ADX e DI"[14, 2] c1 = (mioADXR > 50) c2 = (mioDI > 0) IF c1 AND c2 THEN BUY 3000 CASH AT average[x,tipomm](low) limit ENDIF // Exiting long IF longonmarket THEN SELL AT average[x,tipomm](high) limit ENDIF set stop %loss 15

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

  //tipomm=2 //x=5 - I normally use a number between 5 and 10 DEFPARAM CumulateOrders = False   // Entering long ignored, ignored, mioDI, ignored, mioADXR, ignored = CALL "PRT - ADX e DI"[14, 2] c1 = (mioADXR > 50) c2 = (mioDI > 0)   IF c1 AND c2 THEN BUY 3000 CASH AT average[x,tipomm](low) limit ENDIF   // Exiting long   IF longonmarket THEN SELL AT average[x,tipomm](high) limit ENDIF set stop %loss 15

   

  1 user thanked author for this post.

  Nicolas

  

  gabri

  Participant

  Topics: 6

  Replies: 116

  Been thanked: 39 times
  11/06/2017 at 11:15 AM #51631
  Nicolas

  Hi Gabri, thanks for your contribution for the benefit of everyone here!
  About this specific strategy, I’d like to know a little more:
  _ what instrument, spread and / or fees?
  _ what timeframe?
  _ optimized settings: did you try WFA?

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20828

  Been thanked: 4742 times
  11/06/2017 at 12:31 PM #51652
  gabri

  Nicolas,

  I use stocks (italian, german and french), daily timeframe and no spread on the strategy I posted. I used this system for several months in real trading (manual) and got always more than 70% winning trades and better gain/loss ratios than the one published. I didn’t WFA.

   

  Gabri

  

  gabri

  Participant

  Topics: 6

  Replies: 116

  Been thanked: 39 times
  11/06/2017 at 1:24 PM #51655
  Nicolas

  Fine!, but how do you define what stocks should be a good candidate to be traded by the strategy?

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20828

  Been thanked: 4742 times
  11/06/2017 at 2:22 PM #51666
  gabri

  I pick stocks with high CSI or ADXR. I take this data from the indicator I posted.

  

  gabri

  Participant

  Topics: 6

  Replies: 116

  Been thanked: 39 times
  11/06/2017 at 2:32 PM #51667
  Nicolas

  I made some backtests and I found it works way better with a major trend filter, since this is a long only strategy with a sorta ‘mean reversion’ theory, what do you think? (trading only if Close>Average[100-200] ..).

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20828

  Been thanked: 4742 times
  11/06/2017 at 4:14 PM #51673
  gabri

  Nicolas,

  I was using it just for small and short trades and I am sure that there are better way to use it. Every input is more than welcome. How would you code it with your system?

  

  gabri

  Participant

  Topics: 6

  Replies: 116

  Been thanked: 39 times
  11/06/2017 at 4:37 PM #51682
  Nicolas

  I coded it this way:

  (trendperiod is an optimized variable), a 100 days period is a good compromise between performance and trades quantity..

  DEFPARAM CumulateOrders = False tipomm=2 x=5 //I normally use a number between 5 and 10 // Entering long ignored, ignored, mioDI, ignored, mioADXR, ignored = CALL "Mio - ADX e DI"[14, 2] c1 = (mioADXR > 50) c2 = (mioDI > 0) c3 = close>average[trendperiod](close) IF c1 AND c2 and c3 THEN BUY 3000 CASH AT average[x,tipomm](low) limit ENDIF // Exiting long IF longonmarket THEN SELL AT average[x,tipomm](high) limit ENDIF set stop %loss 15

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

  DEFPARAM CumulateOrders = False   tipomm=2 x=5 //I normally use a number between 5 and 10     // Entering long ignored, ignored, mioDI, ignored, mioADXR, ignored = CALL "Mio - ADX e DI"[14, 2] c1 = (mioADXR > 50) c2 = (mioDI > 0) c3 = close>average[trendperiod](close)   IF c1 AND c2 and c3 THEN BUY 3000 CASH AT average[x,tipomm](low) limit ENDIF   // Exiting long   IF longonmarket THEN SELL AT average[x,tipomm](high) limit ENDIF set stop %loss 15

   

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20828

  Been thanked: 4742 times
  11/06/2017 at 7:23 PM #51728
  gabri

  Nicolas,

  I think your addition it’s quite interesting but I personally think that launching a screener first (CSI or ADXR) and then running the code on the first 10-15 stocks of the list (building a portfolio) has better results. I quickly tested your version on a portfolio and seems to be carrying lower profits. I like the fact though that you have better gain per trade.

  Gabriele

  

  gabri

  Participant

  Topics: 6

  Replies: 116

  Been thanked: 39 times
  11/06/2017 at 8:30 PM #51729
  Nicolas

  Ok 👌

  What about the rebalancing of the portfolio made with this kind of stocks pick?

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20828

  Been thanked: 4742 times
  11/06/2017 at 8:43 PM #51732
  gabri

  Good question. I don’t use too much this system anymore but I used to rebalance weekly. I always keep between 20 to 35 stocks in a short trading portfolio. When I sell I am not rushed to find a new stock to replace the old. I just run this screener (or some other screener focused on short term trading) once a week and if I find a good stock I add it.

  

  gabri

  Participant

  Topics: 6

  Replies: 116

  Been thanked: 39 times
• Author
  Posts