Help Needed to Debug Existing EA with proven strategy | Page 33 | 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

  • #641
  • May 2, 2018 10:38am May 2, 2018 10:38am
Quoting Crooky
i know trading the M1 isn't everyone cup of tea however I have perfected it to an art and am very very profitable this EA will take the pressure off me to monitor as I do, Cheers
I just wonder Crooky, if you are indeed very very profitable, why not hire a coder who can code your EA exactly to your own specifications?
Seems 1:1 logic to me.
Persistence will get it. Consistency will keep it.
  • #642
  • May 2, 2018 11:11am May 2, 2018 11:11am
Quoting SwingMan
{quote} 1. The idea with "optimized hours trading" is good, I also did something like that. For Philips system you would probably have to set tight TPs, and he can think about it when he has time - it's his system. 2. In a next version I will do the calculations of the limits with ATR if the current version stays stable
Thanks for replying SwingMan,

I hope Phillip understands the idea behind optimizing trading window, it can make a HUGE difference in some strategies what time of day they are traded. Even if theoretically we can't visualize this, when a optimization is done for all open/close possibilities the results can be surprising (by my experience).
  • #643
  • May 2, 2018 1:08pm May 2, 2018 1:08pm
Quoting danielhsc
{quote} Thanks for replying SwingMan, I hope Phillip understands the idea behind optimizing trading window, it can make a HUGE difference in some strategies what time of day they are traded. Even if theoretically we can't visualize this, when a optimization is done for all open/close possibilities the results can be surprising (by my experience).
Until Philip understands the idea and I program something, maybe you do a quick check and post a result table ...
- Choose first suitable parameters, time Frame M15, possibly make an optimization.
- Start the EA only for three months (Jan-Feb-Mar 2018) and for a duration of 3 hours (e.g. 7-10, 8-11, 9-12, 12-15 and 17-20,18-21)
- For April test the same hours whether they are further ahead or behind in the ranking.

I think that will be an interesting study, and maybe will emerge a good system variant .
  • #644
  • May 2, 2018 3:55pm May 2, 2018 3:55pm
Quoting detector
Swingman, latest test of the Dow, bit surprised no action/trades taken as highlighted, (test dates 2017.04.02 - today) note the orders, looking very manageable, Could I ask others to put up some tests please good or bad, we seem to be hitting stalemate, many comments and suggestions, let us all know what your seeing in testing, thanks again SW for the valuable time your dedicating to the thread, where is Phillip, he works some hours lol, D {image} {image}
Dear detector
I guess spread of 10 is far to low. Try to calculate with spread 100. See what will happen... :-(
  • #645
  • Edited 4:32pm May 2, 2018 4:10pm | Edited 4:32pm
Quoting kofix11
{quote} Dear detector I guess spread of 10 is far to low. Try to calculate with spread 100. See what will happen... :-(
Hmm, I may have it wrong!!! lol, current out of hours is 60, but in play 10, I tried current on EUR/USD and 14 out of hours, but in play 8, hope this helps to determin, D
Edited my post kofix11, could someone clarify please, I'm laughing at myself, as all this time been testing on the spread of 10 for the dow, thanks, D
I don't need to be good at the game, only good enough to beat my opponent,
  • #646
  • May 2, 2018 4:39pm May 2, 2018 4:39pm
hahahah, test dax with low TP and Pip positions + 10 spread, you get a 1000% increase in 1 month xd
Building a Freedom All Time Return: 5,178.8%
  • #647
  • May 2, 2018 4:57pm May 2, 2018 4:57pm
Quoting J1mm
hahahah, test dax with low TP and Pip positions + 10 spread, you get a 1000% increase in 1 month xd
Have you tested Jimm, if so can you post your findings cheers, D
I don't need to be good at the game, only good enough to beat my opponent,
  • #648
  • May 2, 2018 5:20pm May 2, 2018 5:20pm
This is a snapshot taken from my live account, it shows the out of hours spread at 60, this is a little confusing, could someone clear this up, thanks D
Attached Image
I don't need to be good at the game, only good enough to beat my opponent,
  • #649
  • May 3, 2018 1:33am May 3, 2018 1:33am
Hello SwingMan,

I tested the 4.2 ea yesterday and it was very stable, but today when my mt4 restarts, the ea sends a bunch of orders again reaching the limit orders.

So, i will test the 4.2.1 now and tell you tomorrow if it's good or not.


Thanks for your work !
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Screen Shot 05-03-18 at 08.31 AM.JPG Size: 840 KB
Click to Enlarge Name: Screen Shot 05-03-18 at 08.34 AM.JPG Size: 881 KB
  • #650
  • May 3, 2018 2:37am May 3, 2018 2:37am
Quoting albert666
Hello SwingMan, I tested the 4.2 ea yesterday and it was very stable, but today when my mt4 restarts, the ea sends a bunch of orders again reaching the limit orders. So, i will test the 4.2.1 now and tell you tomorrow if it's good or not. Thanks for your work ! {image} {image}
Thank you @albert for your test.

It may be that the cause has not been fixed, or I have posted a wrong version (that can happen to me)...

Now I'm just posting a testing version only for you ...

- As parameter we now have "Check_PricePosition" = true or false.

- If the "MagicNumber = 0" and "Check_PricePosition=true" the program set automatically "MagicNumer = 7" and for "false", "MagicNumber = 8"
The EA Comment is also adjusted, so that you can start the EA in two charts and the trades can be tracked separately.
Quote
if(Check_PricePosition==true)
{
sCommentEA=sCommentEA+"Y";
MagicNumber=7;
}
else
{
sCommentEA=sCommentEA+"N";
MagicNumber=8;
}
- The solution I believed to work is this (maybe is good, maybe not...):
Quote
if(Check_PricePosition==true)
{
if(Bid>nextSellPendingEntry) bSend_SELL_OrderAllowed=false;
if(Ask<nextBuyPendingEntry) bSend_BUY_OrderAllowed=false;
}
I will continue to analyze what happens, maybe find something else.
Attached File(s)
File Type: ex4 smForexRM EA_v4.2.2-albert.ex4   65 KB | 535 downloads
  • #651
  • May 3, 2018 3:01am May 3, 2018 3:01am
Quoting SwingMan
{quote} Thank you @albert for your test. It may be that the cause has not been fixed, or I have posted a wrong version (that can happen to me)... Now I'm just posting a testing version only for you ... - As parameter we now have "Check_PricePosition" = true or false. - If the "MagicNumber = 0" and "Check_PricePosition=true" the program set automatically "MagicNumer = 7" and for "false", "MagicNumber = 8" The EA Comment is also adjusted, so that you can start the EA in two charts and the trades can be tracked separately. {quote} - The...

Thanks for the update SwingMan,

I test the 2 eas (4.2.1 and 4.2.2-albert), they are set and working.

We will see tomorrow
  • #652
  • May 3, 2018 3:07am May 3, 2018 3:07am
Quoting kasrafx
{quote} Hi kamtal, May I know your EA setting and time frame at which you traded? green pips to you
Hi Mate,
I'm using the original tictac version as I am not using the Hybrid additions that were made ro later versions. So far so good this week will post weeks results at week end.
settings screenshot below
Attached Image (click to enlarge)
Click to Enlarge Name: settings.png Size: 67 KB
  • #653
  • May 3, 2018 3:08am May 3, 2018 3:08am
Quoting albert666
..., I test the 2 eas (4.2.1 and 4.2.2-albert), they are set and working. We will see tomorrow
Version 4.2.1 is identical with the Parameter "Check_PricePosition=false"
If you test on the same platform, is better to set on 4.2.1 a MagicNumber other as the default 7.
  • #654
  • May 3, 2018 3:21am May 3, 2018 3:21am
Quoting SwingMan
{quote} Version 4.2.1 is identical with the Parameter "Check_PricePosition=false" If you test on the same platform, is better to set on 4.2.1 a MagicNumber other as the default 7.
Ok ok, no problem, thanks !
  • #655
  • May 3, 2018 6:07am May 3, 2018 6:07am
Yop SwingMan,

The version 4.2.2, is easily your best version.

The expert have less less trades than others before, so more secure.

It's the only one who survives a 2 years backtest (2016-2018).
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Screen Shot 05-03-18 at 01.05 PM 001.JPG Size: 238 KB
Click to Enlarge Name: Screen Shot 05-03-18 at 01.05 PM.JPG Size: 316 KB
  • #656
  • Edited 6:34am May 3, 2018 6:12am | Edited 6:34am
Here is a comparative between the 4.2.1 and the 4.2.2, on a backtest from first january 2018 to today.

The first picture, is 4.2.1 and look at the total number of trades : 1349.

The second picture, is the 4.2.2, and look at the total number of trades : 504.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Screen Shot 05-03-18 at 01.09 PM.JPG Size: 260 KB
Click to Enlarge Name: Screen Shot 05-03-18 at 01.12 PM.JPG Size: 261 KB
  • #657
  • May 3, 2018 6:55am May 3, 2018 6:55am
Quoting Chake
{quote} I just wonder Crooky, if you are indeed very very profitable, why not hire a coder who can code your EA exactly to your own specifications? Seems 1:1 logic to me.
The indicators for M1 trading are numerous, I am not a coder but I have a half IT degree and from my meagre understanding an EA for my M1 scalping would be of horrendous size - BUT I am open to suggestions
There are many variable due to the M1 volatility wrt time compared to say H1 trading
Its like H1 trading on extreme steroids and is very mentally draining
The more I learn the more I don't know
  • #658
  • May 3, 2018 7:05am May 3, 2018 7:05am
Quoting albert666
Here is a comparative between the 4.2.1 and the 4.2.2, on a backtest from first january 2018 to today. The first picture, is 4.2.1 and look at the total number of trades : 1349. The second picture, is the 4.2.2, and look at the total number of trades : 504.
Is OK.
And now check please my remark for your private version 4.2.2:
Quote
Version 4.2.1 is identical with the Parameter "Check_PricePosition=false"

In 4.2.1 we don't check the price position, and we have more trades.

In 4.2.2 we check the price position, we reduce the trades, and what is important, is that I suppose we don't have anymore false orders. The only problem remain that we can say this only testing with live prices.

I think I find also an other program solution, but is important to know if this is the cause.
(Maybe the use of pending orders, or global variables can be a solution - I must think about.)
  • #659
  • May 3, 2018 7:35am May 3, 2018 7:35am
Quoting albert666
Yop SwingMan, The version 4.2.2, is easily your best version. The expert have less less trades than others before, so more secure. It's the only one who survives a 2 years backtest (2016-2018). {image} {image}
Albert666

I cant get those results for back testing over 2017. can you share settings please. I get near 90% Drawdown.
  • #660
  • May 3, 2018 7:49am May 3, 2018 7:49am
Quoting nihilist23uk
{quote} Albert666 I cant get those results for back testing over 2017. can you share settings please. I get near 90% Drawdown.

Here you go. Change the .txt to .set
Attached File(s)
File Type: txt SwingMan Forex RM 4.2.2-albert - GBPJPY 15 min.txt   2 KB | 412 downloads
0 traders viewing now