Expert Advisor: Did you create your own successful EA? | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

lets share / create expert advisor strategy from indicator 25 replies

How to create an indicator from an expert advisor 3 replies

How do I learn how to make my own Expert Advisor? 1 reply

Poll: Where did you find your very own system/strategy? 2 replies

Can someone create for me an easy Expert Advisor? 0 replies

Attachments: Expert Advisor: Did you create your own successful EA?
Exit Attachments

Expert Advisor: Did you create your own successful EA?

  • #1
  • First Post: Feb 7, 2014 2:44pm Feb 7, 2014 2:44pm
Hello all,

after a long jouney through the universe of EAs, testing probably almost 1000 and losing money on live accounts I am finally planing to devise my own expert advisor.
I will not program it myself but having it programmed.
Now I'd like to know what your experiences with EAs in general and and even more interesting with self devised EAs are.
Either I did not test the existing EAs properly or there is no free good working EA available. I think if there are working EAs why would the people sell them instead of earning money just with the EA itself?
Do you have an self devised EA that is working? => Not asking to make it public here, just asking out of curiosity and hopefully a confirmation that it is really possible to devise steady good EAs.

Greetings,
Myrmica
  • #2
  • Feb 7, 2014 5:49pm Feb 7, 2014 5:49pm
Auto trading is best way to speculate these markets, do share some nice EA if you have one.
I am smart than average trader, I simply copy the best.
  • #3
  • Feb 7, 2014 6:03pm Feb 7, 2014 6:03pm
Quoting I-Robo
Auto trading is best way to speculate these markets, do share some nice EA if you have one.
Do you have some experience yourself with EAs?
At the moment I am still not sure if it is really the best way since I never consitently won with EAs.
  • #4
  • Feb 7, 2014 6:17pm Feb 7, 2014 6:17pm
Quoting Myrmica
{quote} Do you have some experience yourself with EAs? At the moment I am still not sure if it is really the best way since I never consitently won with EAs.
yes my over all experience is good with EA, and I think if some one is looking for consistent profit then only answer of automated trading.

I suggest do not over leverage any account, and use some decent money management, and almost all EA will give positive results, specially the one which are well established in market. even simple MA cross EA can give good results if money management is wisely used.

most people lose because of high leverages, and then they blew because they cant handle the DD.
I am smart than average trader, I simply copy the best.
  • #5
  • Feb 7, 2014 7:52pm Feb 7, 2014 7:52pm
Quoting Myrmica
Hello all, after a long jouney through the universe of EAs, testing probably almost 1000 and losing money on live accounts I am finally planing to devise my own expert advisor. I will not program it myself but having it programmed. Now I'd like to know what your experiences with EAs in general and and even more interesting with self devised EAs are. Either I did not test the existing EAs properly or there is no free good working EA available. I think if there are working EAs why would the people sell them instead of earning money just with the EA...
Hi myrmica - For the past two months I've been working on converting my manual trading strategy into an EA. My intention is not to run the EA 24/7...I'd only like to flip it on temporarily...say during London while I am asleep...or while Im away from the comp for an extended period of time... so that it can open and manage a trade just in case the right conditions are met.

Knowing that I have had relatively good success with my manual trading strategy, I was also just curious as to what kind of results I could generate running the EA in continuous mode throughout many days/months. Im not completely done coding it...needs some tweaks still...but so far the results have been very promising.
  • #6
  • Feb 8, 2014 2:51am Feb 8, 2014 2:51am
Hello NorthShore,


it's always good if you can program and convert your strategy yourself into an EA.
Let us know how it works if you let your EA run in complete autopilot. What timeframe do you use for your EA?

I am asking because my impression is that backtests get all the more unreliable the shorter the timeframe is. So many more factors come in when its a forward test to change the backtest or demotest results.

Greetings,
Myrmica
  • #7
  • Edited 9:24am Feb 8, 2014 9:11am | Edited 9:24am
Attached Image (click to enlarge)
Click to Enlarge Name: vegas gbppy.gif Size: 8 KB
I think knowing how to code an ea is good in that you can actually test your strategy in a faster way to see whether it works in a large samples of historical data. If I have known about ea, it would definitely shorten my learning cycle. Most of the strategies that you find on the net are completely useless when you tested them out of the large historical data. This can save you time and efforts in actually stop wasting time in trying these strategies on live trading.

ea is good if your trading is mostly mechanical and it can be coded. I know some traders are discretionary but there are still some mechanical rules that they follow too. So, it is entirely possible to code an ea to mimic your trading. The most important point is that you take the emotion out of your trading which can hamper your results. You can also better manage your position by pre-calculating your risk percent by the ea. And when you have tested your ea over a large samples of data, it kinda gives you confidence that your strategy has an edge if the backtesting is positive. By testing and backtesting many strategies, it helps me to focus on important thing in trading and weed out those tricks (esp. those with custom indicators) which had no edge at all. I first started by paying mql programmer to code my ea. Once I am familiar with the mql program, it is easy to code my own ea (though my previous programming knowledge helps). It really open my eyes as I can test many strategies over a short period of time. That is how I manage to find one profitable strategy after one year of testing many unprofitable ones. Now, I am testing it live using an demo account to see its performance in real trading condition. And please don't trade your ea with real money until it proves itself in demo trading. I have learned the hard way

And the other important point is that ea does not sleep So, it can help you to catch a trade while you are sleeping.

Attached is my test results over 6 years data and from an initial deposit of 10,000. It looks positive and I am testing it live to see how it works in real trading condition.
Make everything as simple as possible, but not simpler.
  • #8
  • Feb 8, 2014 12:45pm Feb 8, 2014 12:45pm
Quoting hunting pips
{image}I think knowing how to code an ea is good in that you can actually test your strategy in a faster way to see whether it works in a large samples of historical data. If I have known about ea, it would definitely shorten my learning cycle. Most of the strategies that you find on the net are completely useless when you tested them out of the large historical data. This can save you time and efforts in actually stop wasting time in trying these strategies on live trading. ea is good if your trading is mostly mechanical and it can be coded. I...

nice eq curve you got there huntingpips. can you run a backtest and show the results for only 2013 (as opposed to a full 6 years of data)? starting with a 10k balance, it would be interesting to see the performance.
  • #9
  • Feb 8, 2014 4:31pm Feb 8, 2014 4:31pm
Hi huntingpips,

if the equity curve is only half that good in live forward testing the EA is still superb.
Let us know how it develops.
  • #10
  • Feb 8, 2014 6:05pm Feb 8, 2014 6:05pm
A work in progress, but here are the results so far for last year.

Jan 1, 2013 to Dec 31, 2013

Pair: Euro Yen only
Starting balance: 10k
Lot size: 1.5lots, for every trade
RR: 1 to 1 with additional breakeven logic

Attached Image (click to enlarge)
Click to Enlarge Name: backtest1.jpg Size: 149 KB
  • #11
  • Feb 8, 2014 6:24pm Feb 8, 2014 6:24pm
Hello Traders

Living in NZ time zones etc

Big question do these robots work

If you were trying one out on a live account

you would not want to wake up an see

your account has been emptied out.

With 99% failure rate on the net from people losing on line with money investments etc

What is the best way to attack this.

Can you get a robot to work on micro account with $10K

Can you get a robot that would take small trades,hedging pairs so margins do not blow out.

I tried a couple of things out they did not work at all,I already know how to lose money trading.

The plan is to run the robot on the micro account,just build capital.

What is everybody doing
  • #12
  • Feb 8, 2014 8:11pm Feb 8, 2014 8:11pm
Quoting eliteincome
Hello Traders Living in NZ time zones etc Big question do these robots work If you were trying one out on a live account you would not want to wake up an see your account has been emptied out. With 99% failure rate on the net from people losing on line with money investments etc What is the best way to attack this. Can you get a robot to work on micro account with $10K Can you get a robot that would take small trades,hedging pairs so margins do not blow out. I tried a couple of things out they did not work at all,I already know how to lose money...
I know many live accounts which are running on VPS and they are doing great.
I am smart than average trader, I simply copy the best.
  • #13
  • Feb 9, 2014 12:26am Feb 9, 2014 12:26am
Quoting I-Robo
{quote} I know many live accounts which are running on VPS and they are doing great.
Hello Robo

Thinking about running the E A on micro account,I will check if the broker has the VPS service.

Which Robot are you using

Were do we find out about these robots,I asked a couple people who were selling robots if they use what

they were trying to sell.

Hard case you never hear from them again.
  • #14
  • Feb 9, 2014 3:30am Feb 9, 2014 3:30am
There are no successful EA's.
Forex is a losers game.
  • #15
  • Feb 9, 2014 5:36am Feb 9, 2014 5:36am
Quoting NorthShore
{quote} nice eq curve you got there huntingpips. can you run a backtest and show the results for only 2013 (as opposed to a full 6 years of data)? starting with a 10k balance, it would be interesting to see the performance.
hi NorthShore,

It kinda tank in 2013 (-20%). However, for2007, 2008,2009, 2010,2011 and 2012 (testing for one year period for each year data), it is running profitably. It seems 2013 is bad for my ea for some reason . Any insights for that?

It is back to the drawing board again. haha....more testing again....darn forex :0)

ps: I see your ea is profitable in 2013. What about using risk percent instead of fixed lot for your ea. Maybe it will be better results.
Make everything as simple as possible, but not simpler.
  • #16
  • Feb 9, 2014 3:04pm Feb 9, 2014 3:04pm
Quoting hunting pips
{quote} It seems 2013 is bad for my ea for some reason . Any insights for that? It is back to the drawing board again. haha....more testing again....darn forex
HP - I dont have any insights as to why that is. But I would think its always best to optimize your ea with recent market data. Given that market conditions are constantly changing, I dont really see a need to go back very far with tests.


Quoting hunting pips
{quote} ...I see your ea is profitable in 2013. What about using risk percent instead of fixed lot for your ea. Maybe it will be better results.
Yes Im sure the resulting profits/balance would be substantially higher using a progressive risk increase for position sizes. Problem is that I am a total novice when it comes to coding and luckily for me my strategy is very simple and it didnt take me too long to learn just enough coding to get my EA to where it is right now. I dont know what line of code to use for risk percent....yet....Im sure I could find out though. Im still learning little by little.
  • #17
  • Feb 10, 2014 5:59am Feb 10, 2014 5:59am
Quoting NorthShore
{quote} HP - I dont have any insights as to why that is. But I would think its always best to optimize your ea with recent market data. Given that market conditions are constantly changing, I dont really see a need to go back very far with tests. {quote} Yes Im sure the resulting profits/balance would be substantially higher using a progressive risk increase for position sizes. Problem is that I am a total novice when it comes to coding and luckily for me my strategy is very simple and it didnt take me too long to learn just enough coding to get...
hi, make some correction to my code and now my 2013 is slightly profitable (~15%) with initial deposit 10,000. Make some mistakes regarding the stoploss. Esp. when the mt4 just updated to the new build and cause some rethink for my code. Silly me ;0) If you need the code, I can send to you. But not sure how to do it in this forum.

You can check out http://forum.mql4.com/39420 for some ideas for risk percent. It is good to pay for the first coding job from skilled programmer. I used the frame and re-code to try out many strategies. This is when I found most, if not all, of the strategies that you found online are completely useless. That is why I developed my own idea and code my own ea. I sorta think that it is very risky to share your strategy with skilled programmer, and they may actually use your idea and create one profitable ea. So, use an unprofitable strategy to let a programmer code it first time, and then modify it to suit your own strategy. Once you have a good working code and the framework, it is easier to modify it. One of best investment made as I am able to test many strategies in a faster way.
Attached Image (click to enlarge)
Click to Enlarge Name: eurjpy 2013 best vegas.gif Size: 10 KB
Make everything as simple as possible, but not simpler.
  • #18
  • Feb 11, 2014 1:02pm Feb 11, 2014 1:02pm
Quoting Myrmica
Hello all, after a long jouney through the universe of EAs, testing probably almost 1000 and losing money on live accounts I am finally planing to devise my own expert advisor. I will not program it myself ... Greetings, Myrmica
Manual trading is preferable to any ea based trading and as for EA, most do not work after a few months.
  • #19
  • Feb 12, 2014 2:16pm Feb 12, 2014 2:16pm
Quoting Forexnuts
{quote} Manual trading is preferable to any ea based trading and as for EA, most do not work after a few months.
Not necessarily...a sound strategy (whether its used manually or in an EA) is one based on exploitable, repeatable, patterns. PA based patterns, or Indi based patterns repeat themselves in an infinite loop. Human nature/market nature create those patterns first and once you or your EA can recognize them, you can anticipate the most likely outcome of those patterns. Once you discover the perfect combo of patterns on the right timeframe, on the right currency pair, you can exploit those conditions and collect the pips that occur shortly after those setups. On that note, indicator "triggers" and indicator "patterns" are two very different animals.

Ultimately, the most important aspect of the EA is the money management. Notice the 1.68 profit factor in my backtest. Not really all that great. Similarly, notice that the percent of profitable trades is only slightly past 50%....not all that great either. But look at the end result...500% increase on the account in 1 year...a great return on the initial investment. In summary, I believe it only takes a very slight edge coupled with superb money management to create extraordinary results.
  • #20
  • Feb 20, 2014 9:46pm Feb 20, 2014 9:46pm
hello to all my opinion in this thread will be if you have an extremely profitable strategy and can convert it into ea with built in Money Management then it can give you huge returns... but on the other hand their is a quote that indicators works according market but market don't work according indicator but it follows certain principle without which any market cannot work so if EA is made with the market principles and rules then that may proof to be a profitable EA. Howsoever both the manual traders and the traders who trades with EA should be appreciated for the great efforts and their hard work..

Regards
0 traders viewing now