NAS 2m HULL-SAR trading system

• 10/12/2020 at 6:17 PM #147201
  robertogozzi

  @Matchsolo

  Always use the ‘Insert PRT Code’ button when putting code in your posts to make it easier for others to read.

  Thanks 🙂

   

  

  robertogozzi

  Moderator

  Topics: 69

  Replies: 16455

  Been thanked: 3636 times
  10/12/2020 at 7:01 PM #147207
  nonetheless

  Yes, i’m getting the same error message today although backtests worked fine at the weekend. For now I’m running tests without tick by tick and getting the same results that I got at the weekend with tick by tick checked.

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  10/12/2020 at 8:27 PM #147213
  turame

  I am reassured to note that I am not the only one…

  

  turame

  Participant

  Topics: 26

  Replies: 109

  Been thanked: 12 times
  10/12/2020 at 8:27 PM #147214
  ullle73

  paul, could you send the complete version of yours? 🙂

  

  ullle73

  Participant

  Topics: 51

  Replies: 178

  Been thanked: 7 times
  10/12/2020 at 10:24 PM #147229
  nonetheless

  new version with Paul’s TS and modification to the MaxPositionsAllowed code – thanks to @fifi743

  3 users thanked author for this post.

  fifi743, Paul, boonet

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  10/12/2020 at 11:19 PM #147241
  Paul

  Looks very nice and found more performance! I start doubting too btw how the tick by tick & spread is handled with cumulative positions.

  ullle73, there are too many changes i’am making (layout) and still experimenting.

  About the strategy, i’am taking another approach. The equitycurve is very nice, however it’s depended on multiple positions, which could be a warning? I’am not sure.

  So i’am testing a different way to see if it can work with 1 position, but taking every signal into account. And that works too it seems.

  Nonetheless, try set maxallowedpositions to 1. Something seems off if that’s used in v5

  Here’s an early pic with 1 position.

   

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  10/13/2020 at 8:33 AM #147257
  nonetheless

  this is v5 with max positions =1

  looks ok to me, except that it’s with tick by tick unchecked. the result is consistent with what i was seeing at the weekend though, so seems to me that either this result is correct or none of our results is correct.

  are you getting backtests with t-b-t ?

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  10/13/2020 at 12:59 PM #147290
  deletedaccount051022

  Hi

  Thank you very much for sharing this strategy, it’s a great approach to run a strategy with just a few entry criteria.  I  have been learning to keep entry signals to a minimum, and that the exit and position sizing are as important as the entry signals (highly recommend ‘Quantitative Trading’ by E. Chan, and ‘Building Winning Algorithmic Trading Systems’ by K. Davey).  I ran the results of the latest strategy through a Monte Carlo Simulatro (provided by the author of the second book) which uses a random sampling and replacement approach.  The parameters I’ve used are below, but it’s another way of looking at a strategy to test it’s future robustness.

  Capital: 3,000

  Risk per trade: 5

  As you can see, even at these levels, the risk of ruin is only 4%, provided you can live with a 33.3% drawdown.  Keeping the Risk per trade constant, it would be wiser to run a starting equity of closer to 5,000.

  Regardless, it’s another way of seeing how it may perform in the future, and the results are very strong.  Thank you once again.

   

  1 user thanked author for this post.

  HannesA

  

  deletedaccount051022

  Participant

  Topics: 26

  Replies: 139

  Been thanked: 97 times
  10/13/2020 at 1:15 PM #147292
  deletedaccount051022

  Results from the PRT platform with the same test parameters, for comparison.

  

  deletedaccount051022

  Participant

  Topics: 26

  Replies: 139

  Been thanked: 97 times
  10/14/2020 at 3:35 AM #147361
  Paul

  The test I did was with tbt. Didn’t check the difference with on/off. But I’ve lost my settings (own fault) for that equitycurve. That was without optimised sl/pt & ts.

  In demo strategy, one trade exits on stoploss and positionsize went from 8 to 7. I thought in the live version it only can go from 8 to 0, maybe with exception of exits on stoploss?

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  10/16/2020 at 4:05 PM #147573
  Roger

  nonetheless wrote:

  new version with Paul’s TS and modification to the MaxPositionsAllowed code – thanks to @fifi743

  Thanks for the great work. I ve spotted an error concerning the enter conditions : when you use CountOfPosition when you are short on market, the result is negative so you need to use ABS function. Otherwise the condition is always true and you can sellshort an unlimited number of contracts.

  IF Ctime and c1 and c1a AND C3 and c3a AND C5 and c5a and abs(CountOfPosition) <= MaxPositionsAllowed THEN IF Ctime and c2 and c2a AND C4 and c4a AND C6 and c6a and abs(CountOfPosition) <= MaxPositionsAllowed THEN

  1 2 3

  IF Ctime and c1 and c1a AND C3 and c3a AND C5 and c5a and abs(CountOfPosition) <= MaxPositionsAllowed THEN   IF Ctime and c2 and c2a AND C4 and c4a AND C6 and c6a and abs(CountOfPosition) <= MaxPositionsAllowed THEN

   

  

  Roger

  Participant

  Topics: 9

  Replies: 93

  Been thanked: 30 times
  10/16/2020 at 10:36 PM #147611
  Roger

  Sorry the correct code is :

   

  // Conditions to enter long positions IF Ctime and c1 and c1a AND C3 and c3a AND C5 and c5a and abs(round(CountOfPosition/positionsize)) < MaxPositionsAllowed THEN BUY positionsize CONTRACT AT MARKET SET STOP %LOSS 1.3 SET TARGET %PROFIT 2 ENDIF // Conditions to enter short positions IF Ctime and c2 and c2a AND C4 and c4a AND C6 and c6a and abs(round(CountOfPosition/positionsize)) < MaxPositionsAllowed THEN SELLSHORT positionsize CONTRACT AT MARKET SET STOP %LOSS 1.1 SET TARGET %PROFIT 1.4 ENDIF

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

  // Conditions to enter long positions IF Ctime and c1 and c1a AND C3 and c3a AND C5 and c5a and abs(round(CountOfPosition/positionsize)) < MaxPositionsAllowed THEN BUY positionsize CONTRACT AT MARKET SET STOP %LOSS 1.3 SET TARGET %PROFIT 2 ENDIF // Conditions to enter short positions IF Ctime and c2 and c2a AND C4 and c4a AND C6 and c6a and abs(round(CountOfPosition/positionsize)) < MaxPositionsAllowed THEN SELLSHORT positionsize CONTRACT AT MARKET SET STOP %LOSS 1.1 SET TARGET %PROFIT 1.4 ENDIF

   

  1 user thanked author for this post.

  GraHal

  

  Roger

  Participant

  Topics: 9

  Replies: 93

  Been thanked: 30 times
  10/17/2020 at 10:44 AM #147618
  nonetheless

  Roger wrote:

  I ve spotted an error concerning the enter conditions

  Thanks Roger, you’re absolutely right, it won’t work the way it was. I went for a slightly different solution:

  Once MaxPositionsAllowed = 5*positionsize // Conditions to enter long positions IF Ctime and c1 and c1a AND C3 and c3a AND C5 and c5a and COUNTOFLONGSHARES <= MaxPositionsAllowed THEN BUY positionsize CONTRACT AT MARKET SET STOP %LOSS 1.3 SET TARGET %PROFIT 2.1 ENDIF // Conditions to enter short positions IF Ctime and c2 and c2a AND C4 and c4a AND C6 and c6a and COUNTOFSHORTSHARES <= MaxPositionsAllowed THEN SELLSHORT positionsize CONTRACT AT MARKET SET STOP %LOSS 1 SET TARGET %PROFIT 1.4 ENDIF

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

  Once MaxPositionsAllowed = 5*positionsize   // Conditions to enter long positions IF Ctime and c1 and c1a AND C3 and c3a AND C5 and c5a and COUNTOFLONGSHARES <= MaxPositionsAllowed THEN BUY positionsize CONTRACT AT MARKET SET STOP %LOSS 1.3 SET TARGET %PROFIT 2.1 ENDIF // Conditions to enter short positions IF Ctime and c2 and c2a AND C4 and c4a AND C6 and c6a and COUNTOFSHORTSHARES <= MaxPositionsAllowed THEN SELLSHORT positionsize CONTRACT AT MARKET SET STOP %LOSS 1 SET TARGET %PROFIT 1.4 ENDIF

  This uses a different definition of “Maxpositionsallowed” – not the total number of contracts, but the number of entries regardless of size. This way it will work with MM as the positionsize changes.

  4 users thanked author for this post.

  NickBeam, Paul, GraHal

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  10/18/2020 at 1:29 PM #147665
  nonetheless

  Here’s another variation using 1m, 5m, 15m with the alternative entry conditions (beginning the run when the 5m changes direction, then adding positions on each 1m dip). Goes without saying that it is curve fit to these few months, but could be worth forward testing on demo to see how it plays out.

  5 users thanked author for this post.

  Midlanddave, Roger, swedshare, fifi743, boonet

  

  nonetheless

  Participant

  Topics: 83

  Replies: 1237

  Been thanked: 978 times
  10/18/2020 at 4:04 PM #147668
  fifi743

  Hi nonetheless
  You add this line at the end of the algo and make a test?

  Mrsi=RSI[14](close) if longonmarket and CurrentDayOfWeek=5 and close>positionprice and Mrsi crosses under 70 then sell at market endif if shortonmarket and CurrentDayOfWeek=5 and close<positionprice and Mrsi crosses over 30 then exitshort at market endif

  1 2 3 4 5 6 7

  Mrsi=RSI[14](close) if longonmarket and CurrentDayOfWeek=5 and close>positionprice and Mrsi crosses under 70 then sell at market endif if shortonmarket and CurrentDayOfWeek=5 and close<positionprice and Mrsi crosses over 30 then exitshort at market endif

   

  1 user thanked author for this post.

  boonet

  

  fifi743

  Participant

  Topics: 9

  Replies: 843

  Been thanked: 399 times
• Author
  Posts