Help Needed to Debug Existing EA with proven strategy | Page 13 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Help to debug my EA 2 replies

Developer Needed To Debug my EA 2 replies

MT4 EA coder required to fine-tune/debug existing EA 0 replies

Debug code in strategy tester 1 reply

debug message MT4 EA in Strategy Test 7 replies

Attachments: Help Needed to Debug Existing EA with proven strategy
Exit Attachments

Help Needed to Debug Existing EA with proven strategy

  • #241
  • Apr 8, 2018 3:44am Apr 8, 2018 3:44am
Quoting Philip Lee
Tictac-nur, I think its a bit too early to have incorporated all the extra features into the EA making it a bit too complicated I guess. I think we need to test out thoroughly how well the theory works out in reality first before we start to incorporate all the extra features into it. Tictac-nur, is it possible to code a simplistic hybrid EA trading both buy & sell and will close up all the trades when a preset profit level is hit (generated from all the opened buy & sell orders). The simplistic EA would operate in virtual mode and only have the...
Except I sincerely dont understand what youre saying, the EA TicTac has coded does what youre asking...maybe asking TicTac to explain the parameters of the current algo to you might help..
  • #242
  • Apr 8, 2018 4:10am Apr 8, 2018 4:10am
Quoting Thoughts
{quote} Except I sincerely dont understand what youre saying, the EA TicTac has coded does what youre asking...maybe asking TicTac to explain the parameters of the current algo to you might help..
correct, Tictac-tur ea does exactly what you have just requested and more. Just test it and leave the options you do not want turned off.

Testing last night has showed that gbpjpy may not be the best pair for the ea on fully automatic. Work in progress but spreadbet has already given insight.
  • #243
  • Apr 8, 2018 6:35am Apr 8, 2018 6:35am
Hi guys,

Ya, I did test the EA but could not get it to run in hybrid mode (in trying to back test). When I get home later (still in office now), I will look up to all the screen shots posted here and try to copy the settings there.

Thanks guys.

Cheers.
  • #244
  • Apr 8, 2018 6:37am Apr 8, 2018 6:37am
Can anyone please post their back test results on hybrid mode???
  • #245
  • Apr 8, 2018 6:43am Apr 8, 2018 6:43am
Well done spreadbet - that is the proper way to apply the EA - and I think the only proper way as at this juncture unless we could prove that it truly works in hybrid mode.

Just a word of caution spreadbet, don't blindly turn both buy and sell to true. Do some basic analysis first and decide whether to turn the buyer and or the seller on.

I personally think that at this juncture, the only way that both buyer and seller ought to be traded together is in hybrid mode and not in traded independently from one another.

But good to see that you guys are beginning to see the potential of this strategy and started to explore the possible application.

Keep it up guys.
  • #246
  • Apr 8, 2018 6:47am Apr 8, 2018 6:47am
Tictac-nur,

If what Thoughts & Nihilist23uk stated above is true, ie, the hybrid mode is functioning properly, then please ignore my request above.

I hope I have not caused any troubles to you by my request.

Cheers Tictac-nur!!!
  • #247
  • Apr 8, 2018 7:08am Apr 8, 2018 7:08am
Quoting tictac-nur
Thanks guys.. i found the bugged when running on V2.1 with Hybrid Mode + On_cycle only.. here is the update version.. FOREX RM Version 2.2 DownloadLink : {file} ChangeLog 1) [Fixed] Hybrid Mode + On cycle only 2) [Add] Display Max Total System Drawdown. SO you can monitor what is your actually max drawdown when running on hybrid mode.
Awesome work!! Thank you.
  • #248
  • Apr 8, 2018 8:02am Apr 8, 2018 8:02am
Forex RM V2 01-01-2018 - 8-4-2018
Attached Image (click to enlarge)
Click to Enlarge Name: rm v2.jpg Size: 342 KB
  • #249
  • Apr 8, 2018 8:08am Apr 8, 2018 8:08am
Hi Jamco388,

I presumed that its v2.0 of the EA???

Anyone tested the hybrid mode on v2.1 or 2.2???

Thanks anyway Jamco.
  • #250
  • Apr 8, 2018 8:29am Apr 8, 2018 8:29am
Quoting Philip Lee
Hi Jamco388, I presumed that its v2.0 of the EA??? Anyone tested the hybrid mode on v2.1 or 2.2??? Thanks anyway Jamco.
My findings where Hybrid mode seems to function very poorly in back testing on any setting or cycle etc. So bad I did no keep any records.

Maybe others had better luck than me.
  • #251
  • Apr 8, 2018 8:39am Apr 8, 2018 8:39am
Quoting nihilist23uk
{quote} My findings where Hybrid mode seems to function very poorly in back testing on any setting or cycle etc. So bad I did no keep any records. Maybe others had better luck than me.
Me too
  • #252
  • Apr 8, 2018 8:41am Apr 8, 2018 8:41am
This is one cycle for hybrid mode. 1st pic is from 1-12 january 2018
Attached Image(s) (click to enlarge)
Click to Enlarge Name: 2-12.jpg Size: 132 KB
Click to Enlarge Name: 0115-0205.jpg Size: 135 KB
  • #253
  • Apr 8, 2018 8:49am Apr 8, 2018 8:49am
Quoting jamco388
This is one cycle for hybrid mode. 1st pic is from 1-12 january 2018 {image} {image}
exactly what I found,hangs for days limiting any profit during draw down.seems better to treat them as independent legs.
  • #254
  • Edited 10:42am Apr 8, 2018 8:50am | Edited 10:42am
I'm testing GBPJPY on V 2.2 hybrid and I'm finding better results, but I've only done a couple of months. I'm running tests on the same day on another MT4 using the non-hybrid setting to do direct comparisons on V 2.2.

This is extremely slow. I believe a solution, if possible, would be to have the Hybrid One Cycle setting restart automatically on the next day after the previous trade is closed and let it run full tests for unlimited periods Then proper optimization routines can be done to find alternate settings and improve things so that the edge is found more efficiently. This would be enormously helpful, but I'm not a programmer and I'm just making a suggestion for TicTac-Nur if it is possible.

I know very well that backtests and optimization routines are far from perfect on MT4, but at least we should get a reasonably good indicator of performance and do it reasonably quickly on many pairs to see if we are dealing with something viable.

Thanks again to Phillip and everyone else.

Edited to specify that I'm talking about Hybrid One Cycle continuous backtesting for the change to the EA
  • #255
  • Apr 8, 2018 9:03am Apr 8, 2018 9:03am
Quoting Backspin
I'm testing GBPJPY on V 2.2 hybrid and I'm finding better results, but I've only done a couple of months. I'm running tests on the same day on another MT4 using the non-hybrid setting to do direct comparisons on V 2.2. This is extremely slow. I believe a solution, if possible, would be to have the Hybrid setting restart automatically on the next day after the previous trade is closed and let it run full tests for unlimited periods Then proper optimization routines can be done to find alternate settings and improve things so that the edge is found...
Hi Backspin,
if you get time,just run one month of test to cover Jan 2017 and see what you find. Its a trouble month and if we can find a setting on that pair for that month I think it should be good. I failed to find anything,maybe you will have better luck.
  • #256
  • Apr 8, 2018 9:11am Apr 8, 2018 9:11am
This is an example of optimization results. It has over 2500 lines in the spreadsheet, each with a different set of options that can then be backtested at the user's choice.

I did this test on just one day that had a big loser (6 Feb 2018). I then changed the sub depth to 25 and it solved it (GBPJPY). It's just a quick example and nothing serious, just to show.

For those unfamiliar, when you use MT4 for backtesting, download the crappy MQ data in the History Center function. After the download is done, click the download button again and then it will re-calculate the data to delete many miss-matches. That should bring your accuracy up quite a bit.
Attached Image (click to enlarge)
Click to Enlarge Name: Opti.PNG Size: 125 KB
  • #257
  • Apr 8, 2018 9:17am Apr 8, 2018 9:17am
Please note that the above optimization was only done on one day's data, just for demo purposes.

Here is a pic of the 6 Feb 18 loser I was referring to.
Attached Image (click to enlarge)
Click to Enlarge Name: 6 Feb 18 GBPJPY.PNG Size: 98 KB
  • #258
  • Apr 8, 2018 9:38am Apr 8, 2018 9:38am
Quoting nihilist23uk
{quote} Hi Backspin, if you get time,just run one month of test to cover Jan 2017 and see what you find. Its a trouble month and if we can find a setting on that pair for that month I think it should be good. I failed to find anything,maybe you will have better luck.
Hi Nihilist. Will do but will be a while. I'm distracted in Florida on winter vacation until the 16th, but will have a look there also.
  • #259
  • Apr 8, 2018 9:45am Apr 8, 2018 9:45am
Quoting Backspin
{quote} Hi Nihilist. Will do but will be a while. I'm distracted in Florida on winter vacation until the 16th, but will have a look there also.
Enjoy your break mate, and thanks for all the info.
  • #260
  • Apr 8, 2018 10:03am Apr 8, 2018 10:03am
This is the result on the same trade with an increase of the Sub Depth to 25.
It's pretty well guaranteed that most pairs will have a different Sub Depth setting.

I circled a couple of things in the pic to show the increase in accuracy when you hit the download button again after the data download. The 90% accuracy shown is fiction, but it is an improvement from the raw download.
Attached Image (click to enlarge)
Click to Enlarge Name: GJ Solution 6 Feb 18.PNG Size: 95 KB
0 traders viewing now