Do EA's really work? | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

The Really Really Boring Forex System 97 replies

do expert advisors really work 38 replies

Can an automated trading system really work? 17 replies

How does Forex really work? 1 reply

Attachments: Do EA's really work?
Exit Attachments

Do EA's really work?

  • #1
  • First Post: Jul 17, 2007 10:32am Jul 17, 2007 10:32am
Hi all: I am a newbie but I have study and read everything I can get my hands on about forex. Have demoed a couple of systems with limited success. I have been reading about EA's.

My question do they work or they just another part of Forex marketing lures?

I would appreciate anyone's response.

Thanks

Spent1
  • #2
  • Jul 17, 2007 10:47am Jul 17, 2007 10:47am
Quoting Spent1
Hi all: I am a newbie but I have study and read everything I can get my hands on about forex. Have demoed a couple of systems with limited success. I have been reading about EA's.

My question do they work or they just another part of Forex marketing lures?

I would appreciate anyone's response.

Thanks

Spent1
Hi there
EA and other ways of automatic trade, based on a good strategies are the best way of trade, because they simple do what you programme on it, in my opinion some automatic trade fail, because simple reasons, the strategies is not good, and its like hand trading, or because who define it is always change in, normally for bad, and / or don’t give it time to be successfully.
I have now a day 2 EA running on real accounts, very successful, the looser trades I have now a days is because I don’t resist to put my “brilliant” hand on it.
OldSchool
  • #3
  • Jul 17, 2007 10:54am Jul 17, 2007 10:54am
Quoting Spent1
Hi all: I am a newbie but I have study and read everything I can get my hands on about forex. Have demoed a couple of systems with limited success. I have been reading about EA's.

My question do they work or they just another part of Forex marketing lures?

I would appreciate anyone's response.

Thanks

Spent1
I have a friend who would swear on his mother's grave, all his trading captial (which doesnt amount to much), and his pet cat that EA's do not work... BUT I disagree..

I think EA's work, but only work as well as the person programming it can program, and trade. The problem with EA's is people fail to program an ENTIRE trading system into them. For example, some EA's have great entry and exit strategies, but are lacking in money management. If an EA has a good strategy backing it, It should do fine.

The nice thing about EA's is that the psychology side of trading is completely taken out of the equation for success.. However, if you belong to the "all mechanical systems need some discretion" crowd then you would argue against a lot of what I am saying. But there have been tons of traders who have been successful and followed mechanical systems.

Hope this helped
  • #4
  • Jul 17, 2007 11:01am Jul 17, 2007 11:01am
  •  tesla
  • Joined Oct 2006 | Status: Friendly Neighborhood Programmer | 533 Posts
The answer is a bit more complex than yes or no.

An EA is just a program, it does what the programmer coded it to do. In that sense your question is like asking "do programs work?" If you (or someone else) has a strategy with a positive expectancy that can be completely automated, then yes, it can be turned into an EA that will "work".

Something you need to be aware of when it comes to EAs:
It's incredibly easy to throw together a piece-of-crap system and "tune" it to work on historical pricing. I could easily create two EAs a day with historical results that look fantastic, but which would most likely fail in a forward-testing or live-market test. Be very skeptical of back-testing results, especially if they're posted next to a button that says "Buy Now! Only $399."
  • #5
  • Jul 18, 2007 1:54am Jul 18, 2007 1:54am
Thanks for the respomses. Very helpful.

Spent1
  • #6
  • Jul 18, 2007 2:24am Jul 18, 2007 2:24am
Quoting Mtav
Hi there
EA and other ways of automatic trade, based on a good strategies are the best way of trade, because they simple do what you programme on it, in my opinion some automatic trade fail, because simple reasons, the strategies is not good, and its like hand trading, or because who define it is always change in, normally for bad, and / or don’t give it time to be successfully.
I have now a day 2 EA running on real accounts, very successful, the looser trades I have now a days is because I don’t resist to put my “brilliant” hand on it.
Mtav, I'm very curious, how long have your EA's been running?

Did you write them, buy them, or contract someone to write them?

If they are making money consistently, why do you monitor them and sometimes use discretion?

If the only trades that lose are the ones you intervene on, this implies that the EA never loses on a trade, is this what you are saying?

Thanks!
  • #7
  • Jul 18, 2007 3:18am Jul 18, 2007 3:18am
EAs are kinda like Eddie Haskell on "Leave it to Beaver" TV show. Sometimes his nice and sometimes his is not so nice.

I tried EAs and have never succeeded w/o a large drawndown. Bottomline - There no easy path to trading profitably in 4X.
  • #8
  • Jul 18, 2007 3:23am Jul 18, 2007 3:23am
Quoting KudzuFX
EAs are kinda like Eddie Haskell on "Leave it to Beaver" TV show. Sometimes his nice and sometimes his is not so nice.

I tried EAs and have never succeeded w/o a large drawndown. Bottomline - There no easy path to trading profitably in 4X.
Great analogy -- he was nice ONLY when he was trying to suck someone in.....
  • #9
  • Jul 18, 2007 4:07am Jul 18, 2007 4:07am
Quoting blueruby
Mtav, I'm very curious, how long have your EA's been running?

Did you write them, buy them, or contract someone to write them?

If they are making money consistently, why do you monitor them and sometimes use discretion?

If the only trades that lose are the ones you intervene on, this implies that the EA never loses on a trade, is this what you are saying?

Thanks!
Hi there
I start develop my EA, since end 2005, based on program i develop to trade stocks, all programmed by me.
Principal problem is that I have little time for forex, so I have to make programs in order they trade for me.
I have one EA in GBPUSD 1H, and other in GBPUSD 1M, my EA 1H, have made since end of may 15 positives trades and 4 negative, and !M made 38 positives trades and 3 negative trades.
As they are very complex, especially 1M in defining entries and money management, I always making some improve in order to get 95%-100% profitable trades.
As vicious from stocks I’m always give some orders by hand, there where I made lose trades, because I try to guess what happen, my EA on 1H have 5 conditions to make a trade and in !M 12 conditions, some times my human nature makes me to make a trade with only 4 or 10 conditions, can you guess what happen?
OldSchool
  • #10
  • Jul 18, 2007 6:21am Jul 18, 2007 6:21am
It's interesting to consider the possibilities of EA's. I wonder what would happen to the financial community if someone actually wrote an EA that was, say, 90% winner. Yes, would be interesting...
  • #11
  • Jul 18, 2007 6:50am Jul 18, 2007 6:50am
Quoting mtcdavs
It's interesting to consider the possibilities of EA's. I wonder what would happen to the financial community if someone actually wrote an EA that was, say, 90% winner. Yes, would be interesting...
Interesting yes -
but I haven't seen yet an EA that can auto trade for at least 12 months, with less than a 20% drawdown and without ongoing rewriting of the code to accomodate changing market conditions.


.
  • #12
  • Edited 7:14am Jul 18, 2007 6:53am | Edited 7:14am
Quoting KudzuFX
Interesting yes -
but I haven't seen yet an EA that can auto trade for at least 12 months, with less than a 20% drawdown and without ongoing rewriting of the code to accomodate changing market conditions.


.
I wrote this last August and it has been running live ever since..

I think the really good strategies do not get published.

I have not made any adjustments to it since I started it. August 01,2006

These are actual trades, not backtesting or forward testing.

The lots are at a fixed amount. If I used a more agreesive money managment scheme, this would produce even higher numbers... But, I hate risk.... So I keep it ultra conservative. I have, from the beginning used a .20 lot size, or $2 per pip and have not increased it... I may do some sort of percent of stop or account eq in the future.

The bottom line is.... EAs work as well as the person that wrote and executed them. EAs are just another tool and nothing more.
Attached Image (click to enlarge)
Click to Enlarge Name: ScreenHunter_003.jpg Size: 61 KB
  • #13
  • Jul 18, 2007 7:14am Jul 18, 2007 7:14am
I don't mean this in any way like i'm coming onto you, but that is one sweet curve!!
  • #14
  • Jul 18, 2007 7:15am Jul 18, 2007 7:15am
Quoting Blackinc
I don't mean this in any way like i'm coming onto you, but that is one sweet curve!!
Well, I generally do not post this type of post. I am a little embarassed now..
LOL
  • #15
  • Jul 18, 2007 7:24am Jul 18, 2007 7:24am
Quoting smjones
I wrote this last August and it has been running live ever since..
I think the really good strategies do not get published.
I have not made any adjustments to it since I started it. August 01,2006
Nice graph.

The key for me is to complete the year with a healthy profit like you show so far. I do great for the summer or the winter. I adjust lot sizes and use more than one EA to improve the yearly average. To answer the thread topic, EAs do work and make money. So do signal services. The key is your selection and management process. This should not be considered contrary to some of my other posts about the value of Jim's style of trading which I still think is the best pure trading.

Too often EAs are viewed as an easier way to trade. I think creating your EA system is harder and takes more knowledge than normal trading. To achieve success like Scott is showing takes a broad knowledge of the underlying trading principles as well as a knowledge of the EA's code and operation. (Maybe he just got lucky the first time but I never have.) Even if you have someone help you code an EA you still need to understand MetaQuotes and the MT platform well enough to direct your project.
  • #16
  • Jul 18, 2007 7:25am Jul 18, 2007 7:25am
Quoting smjones
I wrote this last August and it has been running live ever since..

I think the really good strategies do not get published.

I have not made any adjustments to it since I started it. August 01,2006

These are actual trades, not backtesting or forward testing.

The lots are at a fixed amount. If I used a more agreesive money managment scheme, this would produce even higher numbers... But, I hate risk.... So I keep it ultra conservative. I have, from the beginning used a .20 lot size, or $2 per pip and have not increased it... I may do some sort of percent of stop or account eq in the future.

The bottom line is.... EAs work as well as the person that wrote and executed them. EAs are just another tool and nothing more.
I agree that the good strategies are not widely known in EA world. For all effort that goes into EAs it seems that very few are profitable over the long term.

Congrats on your success so far.
  • #17
  • Jul 18, 2007 8:03am Jul 18, 2007 8:03am
Quoting nvwine
Nice graph.

The key for me is to complete the year with a healthy profit like you show so far. I do great for the summer or the winter. I adjust lot sizes and use more than one EA to improve the yearly average. To answer the thread topic, EAs do work and make money. So do signal services. The key is your selection and management process. This should not be considered contrary to some of my other posts about the value of Jim's style of trading which I still think is the best pure trading.

Too often EAs are viewed as an easier way to trade. I think creating your EA system is harder and takes more knowledge than normal trading. To achieve success like Scott is showing takes a broad knowledge of the underlying trading principles as well as a knowledge of the EA's code and operation. (Maybe he just got lucky the first time but I never have.) Even if you have someone help you code an EA you still need to understand MetaQuotes and the MT platform well enough to direct your project.
Yes, this is it exactly. There is no short cut for knowledge... I have countless attempts, that did not work out. But, a good EA is built on a good methodology. One of the bonuses of an EA is that if you let it run, it will take all the trades that meet the criteria, not just the ones you think will be good.
If you are trading a purely mechanical system, this is an absolute must, because no one knows what will happen in the future and if your system has an edge it will become apparent.
  • #18
  • Jul 18, 2007 8:09am Jul 18, 2007 8:09am
Maybe you should just leave it to Beaver

Quoting KudzuFX
EAs are kinda like Eddie Haskell on "Leave it to Beaver" TV show. Sometimes his nice and sometimes his is not so nice.

I tried EAs and have never succeeded w/o a large drawndown. Bottomline - There no easy path to trading profitably in 4X.
  • #19
  • Jul 19, 2007 9:28am Jul 19, 2007 9:28am
Some really great incite on this post. I want to truly thank each of you for participating.

Spent1
  • #20
  • Nov 16, 2007 7:31am Nov 16, 2007 7:31am
Quoting tesla
The answer is a bit more complex than yes or no.

An EA is just a program, it does what the programmer coded it to do. In that sense your question is like asking "do programs work?" If you (or someone else) has a strategy with a positive expectancy that can be completely automated, then yes, it can be turned into an EA that will "work".

Something you need to be aware of when it comes to EAs:
It's incredibly easy to throw together a piece-of-crap system and "tune" it to work on historical pricing. I could easily create two EAs a day with historical results that look fantastic, but which would most likely fail in a forward-testing or live-market test. Be very skeptical of back-testing results, especially if they're posted next to a button that says "Buy Now! Only $399."
but using MT4 to test? Even using Metastock and Tradestation?
0 traders viewing now