Performance Strategy Tester using Windows 8 - page 2

New comment
 
Subgenius:


 Feel free to check out my code which this on eurusd 1 year, in strategy tester, takes at-least 1 hour!

Which time period exactly ? Which terminal build ? Default settings ?


1 year, which dates ?

 
angevoyageur:

Which time period exactly ? Which terminal build ? Default settings ?


1 year, which dates ?


1 hour timeframe

2012.01.01 ~ 2013.03.31

Build 445

default settings for sure

 
Subgenius:

1 hour timeframe

2012.01.01 ~ 2013.03.31

Build 445

default settings for sure

Started:

2013.04.01 11:50:45    Heikenashi_FisherTransform__EA inputs: Magic=4; Timeframe=60; MaPeriod=20; MaMethod=1; HAbar=1; Length=6; Price=0; NumBars=0; FTbar=1; Lot=0.1; TrailingStart=30; TrailingStop=5; Loss=30; Profit=100; Retries=3; Slippage=3; Spread=3; tick_timeout=0;


system

Finished :

2013.04.01 12:02:03    2013.03.29 22:59  Tester: order #16 is closed

Result :


I used Metaquotes data freshly downloaded.

 
Subgenius:

1 hour timeframe

2012.01.01 ~ 2013.03.31

Build 445

default settings for sure

I have no doubt that your code could be modified to run very much quicker . . .  for example you call   getHAValue()  on each tick,  but its value only changes when there is a new bar . . .  same applies to  getFisherTransformValue()

I'd consider making these changes and testing the improvement but your coding style means it would take far too much time for me to implement,  so I'll leave it to you to try. 

 
RaptorUK:

I have no doubt that your code could be modified to run very much quicker . . .  for example you call   getHAValue()  on each tick,  but its value only changes when there is a new bar . . .  same applies to  getFisherTransformValue()

I'd consider making these changes and testing the improvement but your coding style means it would take far too much time for me to implement,  so I'll leave it to you to try. 

I gave up testing with your original EA,  it was going to take 3 hours to complete . . .  while I was waiting I revised your code,  fisrt making it a little more readable for me,  then implementing the change I described above,  I also got rid of the string comparisons and made it an int comparison . . .  BUY, SELL and N/A

First impression is that it's around 90 times faster.  2 minutes vs 3 hours. 

 

Modified EA (by RaptorUK) tested in same conditions.

2013.04.01 15:05:47    Heikenashi_FisherTransform__EA_RAPTORUK inputs: Magic=4; Timeframe=60; MaPeriod=20; MaMethod=1; HAbar=1; Length=6; Price=0; NumBars=0; FTbar=1; Lot=0.1; TrailingStart=30; TrailingStop=5; Loss=30; Profit=100; Retries=3; Slippage=3; Spread=3; tick_timeout=0;


2013.04.01 15:06:10    2013.03.29 22:59  Tester: order #16 is closed

23 seconds ! Subgenius, do you still format your system ? :-D
 
angevoyageur:

Modified EA (by RaptorUK) tested in same conditions.

2013.04.01 15:05:47    Heikenashi_FisherTransform__EA_RAPTORUK inputs: Magic=4; Timeframe=60; MaPeriod=20; MaMethod=1; HAbar=1; Length=6; Price=0; NumBars=0; FTbar=1; Lot=0.1; TrailingStart=30; TrailingStop=5; Loss=30; Profit=100; Retries=3; Slippage=3; Spread=3; tick_timeout=0;


2013.04.01 15:06:10    2013.03.29 22:59  Tester: order #16 is closed

23 seconds ! Subgenius, do you still format your system ? :-D

That's interesting . . .  I'm testing with a different date range,  April 2010 to April 2011 also H1 EURUSD  but I have over 350 trades

13:35:35 Heikenashi_FisherTransform__EA inputs: Magic=4; Timeframe=60; MaPeriod=20; MaMethod=1; HAbar=1; Length=6; Price=0; NumBars=0; FTbar=1; Lot=0.1; TrailingStart=30; TrailingStop=5; Loss=30; Profit=100; Retries=3; Slippage=3; Spread=3; tick_timeout=0; 

 
RaptorUK:

That's interesting . . .  I'm testing with a different date range,  April 2010 to April 2011 also H1 EURUSD  but I have over 350 trades

13:35:35 Heikenashi_FisherTransform__EA inputs: Magic=4; Timeframe=60; MaPeriod=20; MaMethod=1; HAbar=1; Length=6; Price=0; NumBars=0; FTbar=1; Lot=0.1; TrailingStart=30; TrailingStop=5; Loss=30; Profit=100; Retries=3; Slippage=3; Spread=3; tick_timeout=0; 

I confirm, for 2010.04.01 to 2011.04.30 I have 396 trades.
 
angevoyageur:
I confirm, for 2010.04.01 to 2011.04.30 I have 396 trades.
Great,  thank you :-)
 

I just let the original version run for the same date range as I ran my modified version for a back to back comparison,  it took the same number of trades, some of the profits/losses were a tiny bit different, I assume due to a difference in spread.

 

Timings:  

Modified version  -   Started  12:58:39  Finished  13:00:19   Duration   1 min 40 seconds

Original version  -  Started  13:35:35   Finished  17:10:46  Duration  3 hrs 35 mins

 

Improvement    130 x quicker

123
New comment