Systems vs. The EA

I have been trading 3 different systems in 3 different MT consoles, with an additional 3 different systems as EA in three other MT consoles. I also do something resembling scalping on Oanda though I consider this entertainment, more than developing a useful skill.

What I have noticed that NO ea I have tested turns a profit. No matter what system I look at, no matter how strictly it follows the rules of the system, it always fails. I have far more confidence in my own MMA based system (largely based on Guppy).

Is this a dead end or is is just that the EA's published are strictly observing rules that are really just not designed for running blind? As a programmer I believe I can emulate the same decisions a person might make providing they are documented correctly.

What are your thoughts on this?

(Please keep in mind that I'm professionally employed to reduce human tasks by using automation, scripts, programming, etc, so for me, its a natural conclusion for me to state that I could program any decision I would make. I've been a programmer for 26 years now)

Trading is an ART - so unless this flexibility is built into EAs they are doomed to fail. Anything based on rigid rules cannot handle the randomness of the market - imho.

regards.

Hi I am a programmer my self and have recently got into forex. I have been working with many EA recently.

I'm also an idiot so don't take me to seriously!!!

You know what most of the EA I have found don't work!! At least when I back test them using Alpari data. I've tried tinkering with stoplosses/take profits which can cause them to improve but again non have shown a profit.

Apart from Fozzy.

So there are EA out there that do work. As for me i'm going to write my own EA. Back test it/ foward test it and then once I am happy with the results run it live.

Hi,

I am not a programmer, but I have good idas about how they should be set up. Altough I have programmed in APT and Fortran for years in the machine tool business, I have not come up with the know how to program with this system. I order for an E.A. to work, you have to go outside the realm of standard rules. Lesser R&R's and that sort of thing. You take chances and benifit from the results. Just my 2 cents

Jim