Topic: Logical Rules


I have noticed that the EA Studio generator uses any possible combination of indicators which match the acceptance criteria, irrespective whether these are associated with long or short positions. Although this is viable in order to meet any acceptance criteria, and provide a nice historical yield curve, for generation of long term trading strategies, I don't see how this would work in practise. What I mean by this is that I see many entry strategies generated where e.g.

CCI crosses downwards
Fast MA crosses downwards
Fast MA crosses upwards & downwards (on different timeframes)

Likewise for exits :

Alligator Lip rises
Stochastic goes above the level line
Williams goes up

Is it possible to add an option where long strategies only use traditional long indicators, ie cross up and never cross down, with the same for exits, ie cross down. And perhaps to ignore duplicate indicators, for example - my entry strategy says Fast MA 20 cross above 30 & Fast MA 15 above 21 - it would be better to just have one MA indicator and let the system find an alternative.

Current behaviour that if I want my trading strategy to exit on a moving average, it is conceivable that the MA might go up for a bar before it goes down, but this is not going to really be a reliable strategy moving forwards. I seem to spend alot of time deleting strategies, even though they match the acceptance criteria, because they are not really intuitive.


2 (edited by hannahis 2021-04-22 12:34:33)

Re: Logical Rules

I totally agree and this issue has been raised several times but some users prefers such behaviours.

And I proposed to have the option to choose.

Ie. Default setting, whatever indicators/conditions allowed

The other option is use only Long "compatible" conditions, i.e. only if it's in accordance with Long trading "theory" not reversal theory for ranging EA.   Because Trending EA don't do the opposite, it only choose compatible trading theory such as Buy when fast cross slow (not the other way round)

I think but grouping indicators into such 2 categories (fast cross slow vs Slow cross fast) we can cut down our generating search by Half (because Generator can ignore the other group of indicators/slow cross fast) and speed up the process.

The 2nd advantage beside cutting down the search by half, the Generator would develop more EA that are Trend compatible and not bias toward mean reversal strategies

Below is the link of my previous post on this subject matter … cess-rate/ raised 3 yrs ago

Re: Logical Rules

Thanks Hannah, glad you agree. I might propose a checkbox on the settings where you can choose to implement that logic or not - Im definitely looking for more trading strategies which are trend oriented and at the moment, the EA studio is only generating half the strategies I want to see. Of course, you can go through each strategy in the collection and only add those that fit your category into the portfolio, but it's time consuming and I am sure the studio would be able to accommodate the different approaches quite easily, it does such a great job of blazing through all possible strats already.

The other option could be to choose which strats are added to the portfolio and have a filter in the collection

Would be good to hear back from Popov on this - Im due to renew my license in a week and this development would be key to making that decision.

4 (edited by Minch 2021-04-29 14:10:54)

Re: Logical Rules

Clearly no answer from Popov, again. This is meant to be the priority area. I have been using EAS for the last year with a paid subscription and have found the experience less than satisfactory - after over a million strategies created and tested, many in a demo account on live data, nothing has worked.

I think the only person who makes money from this platform is Popov to be honest so I won't be renewing my account.

Re: Logical Rules

It always hurt me when someone is leaving sad

6 (edited by hannahis 2021-05-02 06:15:42)

Re: Logical Rules

Minch wrote:

Clearly no answer from Popov, again. This is meant to be the priority area. I have been using EAS for the last year with a paid subscription and have found the experience less than satisfactory - after over a million strategies created and tested, many in a demo account on live data, nothing has worked.

I think the only person who makes money from this platform is Popov to be honest so I won't be renewing my account.

Hi Minch,

I understand your disappointment but Popov has different time lines, date line and priorities and he probably can't meet your dateline requirement within a week to satisfy you.  Further, he really can't be seen pressured by users with such renewal (threat) request or else it will start an endless stream of "threats" and he would be mentally stressed and can't work efficiently.  So at times, it's best to ignore such threats/request.

On your side, if you really can't make EAS work for you despite countless tries.  Then it's about time to rethink your approach or reconsider the usefulness of the product.

I do agree that having some form of Logic rule in EAS generation process would improve our search for profitable EA and you probably can renew your subscription when this feature is finally implemented (if ever).

Probably Popov need to classify his To Do list into 2 basic category 1) Efficiency 2) Effectiveness

He has been making lots of head way in the efficiency category, faster backtest, optimization etc

But sometimes, faster of an ineffective or less effective system, doesn't make it better. 

Having sound trading rules is fundamental and essential in developing robust EA that will endure the test of time and hence increase EAS effectiveness and thus bring about greater success to users in searching for profitable EA. 

If the system/software is bias in generating mean reversal EA that usually can't endure the test of time as compared to Trend EA, then it's really important to reconsider and look for ways to improve EAS by adding rule based logic in the generator (however, do bear in mind there are user(s), I only know 1 user, out there who are happy to have mean reversal EA and thus no complaint with the existing result.  But my hunch is that, many more users (whoever you are, do speak up if you would like this feature be implemented.  Your silence isn't helpful) would hope to be able to control the direction/types of logic rules to be used in the generating process). 

I do hope one day, in the midst of the busyness, a ray of light would dawn upon Popov to relook at the effectiveness of EAS and add in this feature for users so that we have greater control over the kind/types of EA we would like to search.

Meanwhile, let's just hope such features will come and do with whatever limitations we have because despite this (lack of) feature, EAS is still the best tool available for me.

Re: Logical Rules

Popov is doing a great job and just because he does not implement what you are looking for does not mean EA studio is bad or does not work. EA Studio creates whatever YOU tell it to do and in the last 13 years (which includes Popovs previous products), his platforms have been the only ones that crested working strategies for me! Right now "Invesard GBPUSD", an EA that I have created with EA Studio, is number 4 on the mql5 market place, passes a 35 years backtest with an R^2 of 99.48, return drawdown ratio of 70 and not a single losing year and additionally is profitable on almost any other symbol out there without any optimization. It also is profitable on a live account going forward. Without EA Studio, this would not have been possible.

Needless to say that it does not follow your proposed "standard textbook rules" in terms of how the indicators it uses are implemented. I wonder how anyone can suggest such a feature at all, given that 80% of the traders are losing money exactly because they stick to such rules which simply don't work and never have. That is the actual beauty of EA Studio, in that it finds unusual strategies and use of indicators that create profitable strategies in the end!

Keep it up Popov! Thank you!

8 (edited by hannahis 2021-05-02 14:59:51)

Re: Logical Rules

Good for you geektrader that you managed to find success with the current EAS set up.  Does that mean we should stop adding any more features that would enhance other users's success?

What makes you think adding this feature wouldn't make improve other users' search success?

Furthermore this suggested feature doesn't take away the existing effectiveness but adds to it.

EAS is good or bad is really very subjective to users' experience.  If you find success it's good, does that discount other people's "bad" experience? Should we criticize others' who are struggling and wish to have some added features?

Sure, we definitely can fine profitable EA one way or another via EAS but often, it's like searching needle in haystack. 

What make you think adding this feature wouldn't increase it's effectiveness, especially for people who belief the fundamental of trading theory is still essential, tell that to those professional traders who trade manually to throw out the fundamental of trading. Who knows such "unusual" strategies are 1 in a billions (data and combinations) which I believe you have used 35 years of backtesting and lots of searching to get it.

Lastly, to suggest new features doesn't mean Popov is doing a bad job.  If any suggestion is deem as a personal attack on his hard work and excellent work, then we all better be silent and not speak anymore nor suggest anymore.

Face it geektrader, not everyone here are able to find success in EAS, why crash the hope of others if they need some additional tools to help them out and what's wrong in trying to suggest features what would make EAS better?  Unfortunately, EAS needs to appeal to greater masses and EAS has to cater to more than 1 customer

Again, as for me EAS is still the best tool out there (for me and I must also accept that this positive experience does not necessary apply to others, nor will I discount their experience), does that mean I've no more expectations and hope for it and thus should stop suggesting new features?