Non-Profitable EA's & Valuable Information...

So I've been a programmer but new to EA's.. I think its going to be a new obsession of mine.. I would like for anyone to kindly share observations made from viewing EA's that work and dont work etc..

I went through a ton of EA's on FF and its very funny to see all these back test showing horrible results and the correlation between the EA's and why most are losing bring very good insights into why most traders lose and so forth...(don't get me wrong some did profit well but only very very few that i've found.. 90% model accurate of course)

So next week I will program my first EA, it will be my current system and from there I will program new ideas and old systems I traded.. Pretty much find one that works well and always have my latest and best one swapped into the live environment (if it passes back and forward testing respectively).

I really like robot trading because I can rarely be up for london opens and I know I will always be a mechanical trader for the most part, so programming EA's is very convenient... especially if a 24hr trading method is needed.

With that said I'm making the switch to EA trader although I doubt my aspects of being a non robotic trader have much of an advantage..



I would like to open the thread to any related topics.. any EA's you find interesting, best brokers who support mt4, haters.. lol

any suggestions for best broker allowing mt4.. I'm sure its a horrible discussion start but I'm with oanda now and I believe they don't allow mt4.

Thanks

Why is it so hard to make a profitable EA? I would imagine that a trendfollowing EA would work just fine since currencies are supposedly trendihg..

Just dont expect megaprofits with that method. Scalping EAs will probably never work since the brokers have millions ways to mess with those.

Trend detection is one of the hardest challenges in automated trading. Even the best ones cannot come close to a human's ability to identify a trend. The automated ones typically are late to detect the trend and stay in it too long (lagging). I think the only approach that will work is a semi-automated one in which the trader identifies the beginning and the end of the trend and the EA handles the trade management.

Edit: Right behind automated trend determination in terms of difficulty is automated Support and Resistance detection. So being unable to automate both trend and S/R makes a successful automated strategy difficult to implement. This may be why a lot of EAs use indicators instead of price action and are failures.

I actually have my main method which I believe detects trends pretty well without any indicators etc.. but ofcourse my tests will be the determining factor if it works...

It's not a fact if an EA's is profitable more than the fact people take there so called winning strategies.. someone sees its mechanical and programs it.. and the EA shows a whole different story about long term problems..

One of the things I noticed on the EA's is that moving averages and scalping methods were among the worst EA's..

Anyone want to share there experience as to which prove to be horrible or best


Thanks!

Also.. exits with EA's show human error as well..

Many are so entry focused the exits seem to be determine in a way that can allow for millionaire profits if the big run comes.. but it never comes and that means they open themselves to millionaire loses so to speak..


I enjoy looking to the EA's code and back tests to see years of data and logic in a few minutes or hours.. huge learning curve.

I can't believe they moved my thread tho.. this is a discussion about information that can be learned from failing logic of EA's and the ones that do work etc..

I didn't want to have a discussion in platform tech about how to program EA's and junk... arg.. well hopefully they know what there doing moving my thread

I told them to put one of my threads in the trash 3 weeks ago and that never happen but they move this one in a day.. WTF! lol