Opinion - very successfull EA - $3000 account to $6300 in two weeks (could have been $9000) - page 2
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
I can easily spend 10-12 hours at this in a day and make little progress. Yeah, tell me about it :) .... It doesn't take long to realize why people on this forum would not hold your hands .... now does it. Let me be honest with you, IMO, I don't think it's easy for anyone to debug/create programs without taking up much of their time. Sure, I can probably do something which takes you 10-hours in 15-mins now. But it toke me 10-hours to figure it out back then. No way I'm gonna leave something taking me 10-hours to figure out now, to help someone directly. Generally we don't learn the big (why) unless we struggle.
Profit Factor: 1.43 ; Maximal Drawdown: 1 415.00 (33.89%) : Average profit trade: 61.89 loss trade: -45.57. Those are all exceptionally good numbers. It's usually better to look at the Equity during the life-span the trade was open instead of the final results like above. 2-weeks results is usually considered short for evaluating performance. However, there's no consensus on #of-trades or time-span that validate results. Any results back-or-forward tested cannot validate what Will work in future. However, it tells us what worked in Past and didn't work in Past. If we're to choose one of these, which would you choose, IMO the one that worked of course.
The purpose of doing Statistical Analysis is to get characteristics of the system. Also, it's used to evaluate if System-A did better than System-B which is the same question we're trying to answer in the above paragraph. If somehow you don't gain the experience to complete a decent back-test but keep showing results like this for about 3-months, Me /Myself / I/ Personally would consider testing with real money.
The list of references I wanna provide are too extensive so I'll just recommend doing a google search in the format mql4.com + "whatever". Example "mql4.com money management". There's lots of mql4 tools which could help you evaluate a system beyond the standard Report. Btw. I don't wanna give you the impression that I'm a Advance Trader or Coder. I consider myself a newbie as well, no 20 years experience here I'm afraid I barely have 1.
thanks sakis, Im not saying I have found the holy grail, in fact I am questioning that I have missed something as it seems too good. I may have made a mistake somewhere. I have respect for anyone in any industry or profession.
you assume a lot about me and what I think / believe and it is wrong, so please, if you cant answer my question constructively, without reading into it in erroneous ways, with respect, please dont answer at all.
i am sorry for my criticks i dind knew you was asking for advice I hope you get the best
Hi
I am looking for the opinion of some of the experienced people on here. I have programmed a simple EA, and it seems to work fantastic on a demo account.
I know trading live is different, but what I want to know is in what way is it different and could this difference have an effect on my EA performance ?
AND how common is it to create a profitable EA. I was looking at the results of the automated trading competition 2010, and the winner had a final account balance of @$77,000 from a $10,000 start. This is maybe my second month programming EA and by projection my EA would have won that competition, getting to over £1,000,000 in 12 weeks. So how common is it to create an EA that produces good results. The reason I am asking is that it seems too good to be true and I am worried I have missed something.
I can send this weeks account summary to anyone who is interested.
Michael
ps the reaon I say it could have been $9000 is that I missed £880 dollars while messing with the EA program and then today I turned off the MT4 to go to work and came back, and if I had left it on, it would have gone to $7,700 instead of finishing at $6,300.
** Also, this is a (projected) 12875 % increase in 12 weeks - although there are similar results in the automated trading competition results, it seems a bit too much for someone like me who is completely new to programming and MQL
MickGlancy - I don't believe something is too good to be true. If you have found a strategy that works, you have found a strategy that works. Just be sure you understand WHY it worked and you've got something.
If find myself in a similar boat as you. I've been studying the markets for years and have lost lots of money before I found MetaTrader and its incredible ability to test strategies with fake money.
My method is fairly simple. I look at gold prices in various currencies with a 144 period bollinger band on 15 minute charts. This combination seems to balance volatility with price stability fairly well and allows one to make fairly decent predictions on where a price can go. I've watched these charts for about two months now and the price always seems to behave exactly as I want it to, but without a computer controlling my entry and exit points, I'd need to watch the market 24/7. - definitely not an option. Lends itself to emotional trading and besides, I can think of better things to do.
So, price > upper band, wait for a new bar to open and close under the upper band and go short - liquidate when price hits lower band.
Inversely, price < lower band, wait for a new bar to open and close above the lower band and go long - liquidate when price hits upper band.
Now you may say that's a horrible way to trade because a trend can start and kill your position. This is why I use very tiny lots (0.1) with at least $3,000 for margin. This allows me to use a method called "Dollar Cost Averaging" to make sure I can take even a small profit if things go horribly wrong. Each time a buy signal is generated according to the above method, I go short with a new order, liquidating all of them at the same target. Opposite is true for sell signals - go short each time a new signal is generated and liquidate all at the same target.
One thing I've noticed about this strategy is that an entire bar (between open and close) almost NEVER goes back into the bands without hitting the opposite side. And if it does, the dips are more than tolerable.
So far, the EA I'm developing to trade this strategy is only supposed to produce an alert when buying and selling conditions present themselves, but when I run in the strategy tester, I get an alert for EVERY TICK that my program wants to go short, and I can't figure out why. Perhaps you could tell me what I've done wrong?
So far, the EA I'm developing to trade this strategy is only supposed to produce an alert when buying and selling conditions present themselves, but when I run in the strategy tester, I get an alert for EVERY TICK that my program wants to go short, and I can't figure out why. Perhaps you could tell me what I've done wrong?
You need a way do define once per bar. Since you're piggy-backing on this thread as well. Hope you don't mind me showing your back-test as of Jan-March 2010. It looked pretty good to begin with but failed in trending periods. Maybe I'm missing the part about Gold, but I tried to stay as true to your description as I could. Anyways the EA and files are attached. I'd like to see what type of results you generate.
You need a way do define once per bar. Since you're piggy-backing on this thread as well. Hope you don't mind me showing your back-test as of Jan-March 2010. It looked pretty good to begin with but failed in trending periods. Maybe I'm missing the part about Gold, but I tried to stay as true to your description as I could. Anyways the EA and files are attached. I'd like to see what type of results you generate.
Wow - did you do this from Scratch? Thanks! I tried it out and found the problem. I'm starting with $3,000. if you don't make an immediate gain with the "no more trading under $3,000" rule then of course the program will fail. The strategy is designed to take advantage of swings, even if a trend starts. I changed the following lines:
if(AccountEquity()<100){
OrderSelect0(0,'t');Alert("No Trading <$100 for margin");
return(0);}
and it worked fine - steady profits from January until present. I can't post screenshots right now, my cousin's computer won't let me open Paint. That's another thing I have to fix :P
I also changed the value for Lots to AccountEquity()/30000 - a hundredth of a lot for every $300 available margin but that doesn't seem to change the results, oddly enough.
@trivates
I plugged your entry signals into my EA and ran a quick backtest, the results are attached. It seemed to perform better on the M5 rather than M15 but ultimately was not profitable. I did not include a 3000 margin requirement since I tested with a $10K account. Lot size was 0.1. Testing period was the year-to-date. In the results attached is a text file I have my EA spit out upon completion of the backtest which includes additional statistical data. The strategy has some potential, perhaps with some filters and/or money management due to the low number of consecutive losses but the way it losses is by leave you upside down in a trade going horribly wrong when the market begins to trend. I did not use any tp or stoploss.. trades were exited on the reverse entry signal (and a new trade opened instantaneously). I'll pull a few things from my arsenal and play with your strategy some more.. if you have any ideas you want to test feel free to let me know, Ill dont mind running them for you.
Just put another 20% on to another account I was using for testing.
out of 951 possible points this week, I have got 800 of them so far, and that is partly because I had to go out for an hour this afternoon.
Perhaps you could tell me what I've done wrong?
Id love to tell you that but I cant work the backtester, I do everything on demo
AND ive just realised what the journal is for LOL
Thanks Sakis, no hard feelings then eh ? :-)
and thanks Trivates for the positivity, and femubs, the account summary is on page 1 of this thread, and I will be giving the EA away when it works :-) dont worry, but for now it has to stay with me.
Uzben "Generally we don't learn the big (why) unless we struggle.", I think Im a junior member of the club now.... :-)