EAs as part of portfolio set - page 5

To add comments, please log in or register
Sergey Golubev
Moderator
110775
Sergey Golubev  
newdigital:
Statements for all portfolio were updated. Just to inform that I am testing portfolio sets posted on the first post of this thread (EAs, settings, monthly results) https://www.mql5.com/en/forum

Statements updated.

Files:
Sergey Golubev
Moderator
110775
Sergey Golubev  

Statements were updated.

Files:
Sergey Golubev
Moderator
110775
Sergey Golubev  
newdigital:
Statements for all portfolio were updated. Just to inform that I am testing portfolio sets posted on the first post of this thread (EAs, settings, monthly results) https://www.mql5.com/en/forum

Updated statements for all the portfolio sets.

Matt Edmonds
18
Matt Edmonds  

Risk managed portfolio

Hi all,

I have just done some analysis of the systems that have been tested over time here, and reviewed those that have the best risk management as demonstrated by potential drawdown vs gains over the time that they have been live traded. Three systems stand out from this group and appear to be very well suited to being traded as a portfolio:

MA Channel - USD/CHF only

Daytrading 3tf - EUR/USD only

Step EAs - Step MA Expert v 1.45 - GBP/USD only

Given the extraordinary win vs loss record of Daytrading 3tf, I would trade it with a weighting of 5 times the value per pip of the other 2 systems.

Here are the combined weekly pips for all 3 systems run at once for the 36 weeks that they have all been operational (with 5 x pip count for Daytrading 3tf)

Week Weeks with all systems traded concurrently

Pips

0 0

1 462

2 275

3 0

4 300

5 0

6 715

7 0

8 -49

9 7

10 421

11 41

12 272

13 408

14 -156

15 -4

16 -177

17 88

18 59

19 97

20 -100

21 419

22 162

23 90

24 312

25 -87

26 3

27 270

28 341

29 -93

30 60

31 21

32 24

33 146

34 78

35 -180

36 176

Total 4401 pips

An average of 122 pips per week

If traded at $5 per weighted pip (without increasing as the account value increases) on a 10k account this would result in a total gain of $22,005 in a year with max drawdown at the end of any week at -$900, and a max total drawdown from consecutive weekly losses of -$1685.

Really not bad at all....now I just have to wade through all the info on this forum to extract these systems, and get the correct setup for only these pairs. I imagine that could take a while.

I'd be interested to hear of anyone else's experiences with these systems.

Cheers,

Matt

Sergey Golubev
Moderator
110775
Sergey Golubev  

Good evaluation. Thanks.

But there is some ... limitation let's say.

MA Channel is doing good for Alpari rus, North Finence and Fibo Group broker witgh default settings (USD/CHF). It is necessary to change the settings to use it with IBFX for example.

Daytrading 3tf. The same with MA Channel. Because the data of NF, Alpari rus and Fibo Group broker are too different from IBFX.

Step MA Expert v 1.45. The same thing. This EA is very sensitive with broker's data.

Thus, it is necessary to correct the settings of those EAs to use it with particular broker. Little correction may be fine.

Just my experiance.

Matt Edmonds
18
Matt Edmonds  

Brokers

Thanks for your feedback New Digital, much appreciated.

To further complicate matters, my broker is ODL Securities. I may try to get the systems operating on their data, and forward test in a demo account for a while before deciding if it is tracking similarly for results with your own testing.

Could you please advise me the variable parameter settings used in each of the forward tested systems so that I can duplicate them as a starting point.

Thanks,

Matt

Sergey Golubev
Moderator
110775
Sergey Golubev  
mattnz:
Thanks for your feedback New Digital, much appreciated.

To further complicate matters, my broker is ODL Securities. I may try to get the systems operating on their data, and forward test in a demo account for a while before deciding if it is tracking similarly for results with your own testing.

Could you please advise me the variable parameter settings used in each of the forward tested systems so that I can duplicate them as a starting point.

Thanks,

Matt

MAChannel - default settings for H1 timeframe. Other EAs - don't remember now. Check files thread here: it should be EAs with the settings there.

I think you need to optimize the settings for your broker with 90% quality for at least some period of time just to correct the settings. Because as I said above those EAs may perform differently with different brokers. Especially MAChannel and Step 1.45.

How to optimize the settings using your broker's data?

Go to Tools, Options, Charts and place 99999999999999999 in Max bar in history and the same in max bar in chart. Means - any big number. Anyone but it should be very big number.

Open Tools - History Center. Double click on pair you want to test, double click on timeframes to download the data. Never press download button from History Center as you will download some unknown data from somewhere if you press it but you need your broker's data. Then open the charts/timeframes you want to test, right-mouse click and Refresh every chart.

Then make Auto Scroll off (on the top of Metatrader just near Zoom out), open chart M1 chart, press Page UP on your keyboard and hold it (hold Page Up). You will see the chart moving back. It is the data downloading from your broker's server. It take a time. Then do it with every timeframe and every pair.

Then Refresh once again.

And then only you may optimize the setting. Try to backtest first to get the period of backtesting to get 90% quality. After receiving the period (Strategy Tester, use date, from, to) you may optimize the settings for that period of time. Did not try to change the settings completely. Will try to get slightly different settings.

Matt Edmonds
18
Matt Edmonds  

Thanks New Digital

Appreciate your assistance,

I have just looked through the posts to find the correct versions of each system, and have been reviewing how they operate. I have been struggling to find the full trading history of each system for these currency pairs however, and would really appreciate that if it is available.

I am still learning about EAs and optimisation. Thanks for your instructions, I have found other documents with them in also. while I am learning, I am going to try forward testing before optimising my brokers data set (and while I am waiting on their reply to my email asking for their 1 min data).

My biggest issue at the moment is that I don't have a reliable ISP, but will be changing in a month or so.

I have now uploaded the 3 systems to my computer, using the following settings for forward testing as a complete portfolio.

Start value of demo account: $10,000

System 1

MA Channel - Standard settings

USD/CHF only

1 Hour time frame

0.3 lots per trade (the system creates 5 trades at once)

Stop loss at -100

TP at 200

Trailing stop at 50

Max potential loss of 5 failed trades at once is around $1,200 but it never seems to perform that poorly in any week in your forward testing. Here's hoping my data set is similar to yours and performs similarly.

The first trade on this system was entered within a few minutes, short USD/CHF, and I didnt have a chance to change the lot size from 0.1 first, but it will be amended in the future to 0.3.

System 2

StepMA Expert v 1.45 - Standard settings

GBP/USD Only

30 Min time frame

0.5 lots per trade

100 pip stop (max loss per 1.0 lot = 1,000)

Max loss per trade on this system is $500.

System 3

Daytrading 3tf - Slightly changed settings

EUR/USD Only

15 Min time frame

0.5 lots per trade, 5 trades created at once, each with different stop levels

with 20, 30, 40, 50 and 100 pip stops - this is the only changed setting from the one you have tested.

This means the max loss per trade is $100 + $150 + $200+ $250 + $500 = $1,200. I wanted to change the risk management to ensure that while it seems to have a very high chance of success, that the potential gains and losses were evened out a bit. Potential gain if the trade goes the right way = $500. If it still keeps performing with great hit rates, like 80%+ without taking stops out, then it will be a very strong system.

So that is my portfolio. I would be interested to hear any more thoughts/comments on this concept. I'll try to update my progress from time to time.

Cheers,

Matt.

Sergey Golubev
Moderator
110775
Sergey Golubev  

You need 1st post of files thread for the performance/statements:

- old statements are attached to the post;

- new statements are giving as a link (in the end of the post) and updating weekly;

- excel file attached to the post (and updating weekly).

StepMAExpert 1.45 EA with settings https://www.mql5.com/en/forum/173375/page9

Sergey Golubev
Moderator
110775
Sergey Golubev  

Portfolio statements were updated on post #44 https://www.mql5.com/en/forum/174701/page3

To add comments, please log in or register