The concept of using EAs.

[Joseff]

Hi, hope you had a good weekend, and your trading week is going fine. Everyone knows two things: 1. EAs don't work forever. 2. They can be profitable during a short period of time. So why no to use EA this way: you buy new trading system and use it until it pays itself 4-5 times, then you stop using it no matter what, even if your grandma ready to sell her kidney so you can keep trading.

I think this concept can change the understanding of EAs. There are many good EAs out there, but they good for limited time.

Bought - made profit - STOP!

Josef:clover:

 

[Joseff]

Any comments?

 

[the hare]

Its a simple case of mathematical expectancy. how many EA's have you tested, and over what duration ? how many of the EA's you tested returned in excess of 400% ?

 

[Joseff]

Its a simple case of mathematical expectancy. how many EA's have you tested, and over what duration ? how many of the EA's you tested returned in excess of 400% ?

4 EAs. Time - 6 month. Results confirm that would be better to stop using them at some point.

 

[Mata Nui]

I think firstly you need to differentiate between commercial EAs (which are primarily created to give possible short term gains for the trader but looks good and therefore makes money for its creator) and personally coding up a mechanical manual system you have found to be profitable.
You talk about dropping an EA after a few months as it becomes less profitable but i would look at re-optimizing first. Market conditions change throughout the year so things like TP and SL should also change.
I like the idea of automated trading. I hate finding i have missed a good entry because i had to do something else. I have coded lots of robots based on manual strategies i have found and had some success with, but have yet to find one that works when coded into a fully automated EA. In fact most i discard after the first weeks forward testing. You may say a week of forward testing is nowhere near long enough. I don’t expect it to make a lot of money in that week. All i look for in the first week is how well it compares to the backtest results. If for example i had a string of a dozen losses on forward test and in several months backtest the worst case was 4 or 5 losses in a row i might take the approach that the backtest was seriously flawed and the EA is not likely to be profitable. I know a backtest cannot replicate forward testing but i do think that you only need a short period of forward testing to be able to determine if your backtest was completely invalid.

 

[Joseff]

I think firstly you need to differentiate between commercial EAs (which are primarily created to give possible short term gains for the trader but looks good and therefore makes money for its creator) and personally coding up a mechanical manual system you have found to be profitable.

Most likely, I'm talking about commercial EAs as I'm not experienced enough to code something myself. So I'm ready to give some money for working and well running trading system.

 

[Joseff]

I still believe this will work for commercial EAs. Paradox: everybody agrees that EAs doesn't work forever. At the same time nobody wants to stop trading with particular EA.

 

[the hare]

I still believe this will work for commercial EAs. Paradox: everybody agrees that EAs doesn't work forever. At the same time nobody wants to stop trading with particular EA.

The most popular of the commercial EA's simply didnt work in the first instance, the asymetric reward|risk just fooled people over the short term into thinking that they did.

What you are proposing will work, but why make it any more difficult or riskier, or even more expensive than it needs to be ?

 

[intradaybill]

This is the fallacy:

You buy 4 EA's for exampel and you trade for time T to make 4X. Then you drop them. Someone else, you buys them right after you drop at time T+ them and loses money. That someone, can be you.

 

[arbus]

There is a reason they are called advisers and not ATM's. Using ea's to open trades and your brain to close them can work very well. Close the losers quick, let the winners run.
The trick is just to never sleep.

 

[Joseff]

There is a reason they are called advisers and not ATM's. Using ea's to open trades and your brain to close them can work very well. Close the losers quick, let the winners run.
The trick is just to never sleep.

Yes! EA isn't ATM!