EA Question for smarties

New comment
 

Is it possible to make an EA, that at a set time will run the mt4 optimizer and

then update its parameters with a new set of optimized parameters?

 

Its possible to make an EA that works fine for the last 3 years without change in parameters.

Yeah, its not a scalper.

 

no its not a scalper i was planning to use the artificial intelligence EA and it has 4 variables but i want to apply the optimization from the last 3 weeks to tomorrow. and i wanted to know if it is possible to make an ea that can run the optimizer and update the parameters? a simple yes, no. would be much apprieciated.

 

in other words is automatic walk-forward optimization a possibility in mt4?

 

why do i even bother asking

 

It is possible. But you don't know what is mql4 language, do you?

Lets make it simple - for the EA to run optimization automatically, you should make a car that makes gasoline and uses it to run

In other words its possible but it requires lots of effort.

Besides few that know -how- to do that also know -why- nobody wants to do it.

Want to know why?

 

In my opinion it is not possible without writing a lot of code outside MT4 but maybe it is just because I don't know how to do it.

To have the final word you should ask Metaquotes: in their Forum their technical guys will give you the final answer forum.mql4.com

 

go ahead dont hold back on me. enlighten me. i think i can program it. it will just take me a wile. i want to test this for my own curious reasons. maybe it wont be profitable but i see something here that many ppl dont. alright.

there is inteiligence in the perceptron algorithm. im just trying to bring it out. even if i cant with all my optimizations its still there i just have to find a way to it.

 

thank u jlpi for the suggestion. i will ask there.

 

The problem is not with the code. You don't need external programs to optimize it. You don't need to run optimization wizard either - it can all be done within the EA itself.

Don't run to their forum either - that was discussed numerous times already (how can an MQL4 programmer not know of MQL4.com? ) but no solution was found.

The problem is not with optimization. You can run it manually.

The problem is with the result of the optimization. Or rather with data you use for it.

There are a couple of scenarios you need to optimize to get results:

1. Trending market - high volatility.

2. Ranging market - high volatility.

3. Trending market - low volatility.

4. Ranging market - low volatility.

5. No volatility.

How will you know what will be tomorrow? How will you identify which setting you should use or what period of the past should be used?

The problem is you need optimized settings for one of the scenarios. Not everyday optimization. Its the same as having no stop-loss. You'll be kicked out of the market sooner or later and become a member of 95% who lose.

Its the problem of not knowing what should you expect tomorrow. If you use settings for ranging market - low volatility, you'll end up with loss if tomorrow happens to be trending market - high volatility.

Besides if you optimize a period with different kinds of market conditions you'll end up either with a system that works no matter what the market brings or with a total loser. For scalpers its almost always the second.

Be careful if you are up to it. It would be wiser to find a way to determine market conditions for the near future and wait for a confirmation and then manually set all settings for a specific scenario rather then optimize ranging market with low volatility which lasted 2 last weeks together with trending market of the last week. In this case you'll have a system that is either optimized for ranging market or simply no profitable results.

I hope you get my point. If you have a goal, work for it and don't let anyone stand in your way.

 

Yes

I understand your frustration and you want a straight answer. I do recommend you move forward with a couple previous suggestions and go to metaquotes.net. As somebody like Slawa and check previous posts in the forum archive as chances are somebody has asked the question before you.

On a separate note, I looked at doing what you want a over two years back, only I had about a couple dozen variables that would 'learn' over time. In the end I gave it up for a couple reasons.

1) If your system requires that much optimizing, IT IS NOT robust and along with this -

2) Yesterdays market performance wonderfully optimized into EA right this instant is NOT a guarantee of what is going to happen even 1 minute into the future.

Yes, what you want can be done but it's very complex and I would suggest you turn all your efforts into building a trading system that is robust no matter what the market throws at it. Your optimizing EA will not give you that as it does not account for volatility of 1) News events, each of which are uniquely fingerprinted and 2) Disaster type situations (9/11) etc.

The only thing that will save you during these periods are, SL's and good MM.

All you need is a 60% accurate system, practice good MM and ensure your TP's are larger than your average SL. It does not get any simpler.

Archer

12
New comment
Reason: