Topic: Maximum Equity Drawdown

I cannot find the "Maximum Equity Drawdown" in the Sorting possibilities (8) of the Repository Collections.
Is it possible to add, because I rather look for small drawdown than e.g. "Net Balance"?

Re: Maximum Equity Drawdown

Anyone have an opinion to this point?

3 (edited by pradeepgolfer 2016-03-08 16:38:31)

Re: Maximum Equity Drawdown

bru1 wrote:

Anyone have an opinion to this point?

you can filter it in the acceptance criteria before the strategies enter the repository

Re: Maximum Equity Drawdown

Yes I know you can filter it together with other criteria, but you cannot sort the results according Equity drawdown, e.g. if there are many various results. Or do I miss something?

Re: Maximum Equity Drawdown

bru1 wrote:

Yes I know you can filter it together with other criteria, but you cannot sort the results according Equity drawdown, e.g. if there are many various results. Or do I miss something?


I do not think you are missing anything.....i could not find it also....but Miroslav is the best person to answer your question

Re: Maximum Equity Drawdown

It is a good suggestion. I may implement it soon.

Re: Maximum Equity Drawdown

Thank you

Re: Maximum Equity Drawdown

Popov wrote:

It is a good suggestion. I may implement it soon.

Could you also please add "maximum stagnation" to the sorting feature?

do or do not there is no try

9

Re: Maximum Equity Drawdown

Do you mean minimum stagnation?

10 (edited by yonkuro 2016-03-09 23:11:44)

Re: Maximum Equity Drawdown

GD wrote:

Do you mean minimum stagnation?

I think the acceptance criteria said "maximum stagnation " right? but as we usually slide the little square button from less to more, I think it's okay to call it "minimum stagnation".

That feature'd be really helpful for me, because I prefer a strategy with less stagnation period.

do or do not there is no try

Re: Maximum Equity Drawdown

geektrader wrote:

Thanks, but I don´t want to use it as an acceptance criteria nor being able to sort already generated strategies by it. I want the actual generation process to use this as it´s goal already. There is a different if just storting strategies by that value AFTER they have been generated with "maximize net profit" as a goal than to directly generate them with "maximize custom fitness function" instead of "NetProfit" or any other of the available generating goals.

I want to highlight Geektrader's emphasis on his above quote.

As what Geektrader requested, he wants it in the Sieving process (the criteria used for generating EA), not the criteria use during sorting process only after the EA is generated. 

Request for Sieving or Sorting Criteria?
There is a GREAT difference between sieving and sorting.  I hope we look into the sieving methods and the sorting will then become more fruitful.  All those request for stagnation, equity, draw down etc...we need to be clear whether we are looking at these criteria in the Sieving Process or the Sorting Process?

What's FSB's criteria used for Generator and Optimizer?

Is it purely a mathematical probability selection or A Profit Hunter?

From a layman's point of view (because I don't know what's FSB Generator criteria or logic behind it's working), I think a purely mathematical probability selection generator, would simply generate EA as long as the sum adds up to show an overall profitability per se.  If this is the main criteria used, we definitely going to have endless number of combinations.  These EA generated are purely random combination that happened to show profit, with no trading theoretical soundness to it.  That's why sometimes, I noticed the combination of conditions used (both open/close) goes against "trading logic".  That's why in the past, I never rely "solely" on Generator or Optimizer to develop my EA (But now, I would use it differently...I'll explain later).


When you are lost, do you just want a compass or a navigator?

A compass is purely just a device, pointing north.  If you have a map and know where you are going, a compass can be helpful, but you are the main person doing all the decision making such as reading the map, pinpoint where you are and where you are going and find the direction, etc.  Your ability to find your way, depends solely on yourself and how much information you have to pinpoint your location, if you are in a foreign country...in a land of forex with multiple and fluctuating information...finding your way on your own is tough.

A navigator has more than the function of a compass.  You input your locations and where you Are, and where you wanted to go and the navigator directs you from point A to point B.  If you are lost, the Navigator can locate where you are with GPS.  Hence, it has a "Locator" function in addition to your compass.

If a Generator is utilizing purely on mathematical calculation to search for all the probably combinations that yield profits, naturally it can generate endless combinations and probabilities till end of the world, the Generator will keep on working.  But the quality/majority of these combinations yield little statistical significance.  And then we thought that the success of finding a profitable EA lies in a) how long/many we can generate and b) our sorting methods; if only we use better sorting methods, we can search/filter/sort out some "gems" out of these millions of combinations (tough luck).

If I have Popov's talent in programming skills...I would like to make the Generator as a Profit Hunter, much like a navigator (looking for profit from point A (opening conditions) to Point B (closing conditions).  Hence, the Generator is not simply a mathematical calculator device but a profit locator device.

I would like the Generator to locate/search for the highest peak/entry position/point A (within certain time frame period, i.e. H4 and any period I would like to choose, depending on what type of EA I want to find, a scalper or a swing trade style) and let the Generator locate/search for the best exit location/close position/point B (the lowest trough) and find what type of combination conditions (both open and close) would fit nearest to these two locations.  Thereafter, I can use whatever sorting process I would like, the chances are, the EA generated would have much more significant results that searching for a profitable EA would not rest so much on the sorting method but the success of finding profitable EA is in the Sieving method. 

Another method would be like the way we use multi market/symbols, we let the generator locate all the multi highest peaks and multi lowest trough and "study" these position and see which combination of opening/closing conditions fit into all these "multi peaks" and multi troughs".  In this method, we are looking not only for a set of combinations that fit only for one peak and trough but we are looking for a set of combinations that yield high probability of fitting into multi peaks and multi trough, so that it is not a issue of random probability but a consistent capability to locate these peaks and trough significantly.


I wonder how many of us know what are the criteria used by the Generator to generate EA.  We put so much emphasis on how to improve the sorting process, that we didn't stop to ask, "What's are the criteria used for the sieving process, in the first place?"  Why not improve the Sieving process first and then work on the sorting process later?

Why I think my suggestion for "Location/Profit Hunter" Generator would work better than the current Generator?

A while ago, I began my Generator "adventure/project".  I want to see whether is there any method I could use to increase the "quality" of the EA generated.

So I input my own conditions (using my trading theory "2 Pillars" concept) into the strategy and let the Generator locate other combinations to add in and using Profit as the selection criteria, eg. Trailing Stop limit.

What's was my rational for the above experiment?

1.  I'm trying to "make" the Generator act more like a locator by giving the Generator some starting point (my own opening conditions) and an ending point with trailing and profit taking position.

2.  I'm trying to reduce the millions of combinations generated by reducing the amount of random probability and hence, also increase the statistical significance of the EA generated.  With each additional opening conditions input, I am giving very specific "direction" to the Generator to narrow it's search for probabilities that fit only into my opening conditions.  And since my opening inputs (are based on sound trading theory), naturally, the EA generated would have greater statistical significance.  Because the Generator is no longer using purely mathematical probability to randomly search for profitable EA per se, but now searching only for profitable EA that are theoretically sound, that fits into a certain trading theory concept (like the 2 Pillars, under Hannah's trading tips).

As the result of this Generator experiment, I managed to have a few good and profitable EA (I could have more if I spent more time in this project) and generating profitable EA from a Generator becomes much more fruitful and yield better results, it's like a trader's dream come true (generating profitable EA is no longer an issue of luck).  From this Generator experiment, I get to see the prowess of FSB's Generator and know it has tremendous potential and capabilities.  Hence, that is why I want to emphasize on improving the Sieving methods, i.e. Looking at the criteria used for Generating EA before we look into ways of improving the sorting process.

Before I want to invest my time and efforts to generate thousands upon thousands of EA via the Generator, I must first understand what's the criteria employed in generating these EA.  I don't want to blindly have faith in the Generator and keep working on it, run it endlessly on my computer and keep sorting and sorting...and yet doesn't even know whether the criteria used is worth believing for.  If it is purely a mathematical probability, I would have better results working on my own theory than to rely on the Generator.  Now I understand the Generator's "limitations", I would work around it, providing the "theoretical" input to transform it from a compass into a navigator.  With this "improvement", I must say, the Generator is a very powerful machine (I wished on one hand, I "discover" how to improve the generator's results earlier and would have taken me less time to find profitable EA. 

Searching for Profit vs Searching for trading knowledge.
But I didn't lose anything either, this "longer" path has greatly sharpen my trading knowledge and theory as a result, which is invaluable too.  If I have taken the "easier path", I would have also short changed my own learning curve and knowledge/trading skills).  Yes, it would be far easier to search for profitable EA using the Generator with my trading concepts input.  Finding profitable EA doesn't necessary translate into having better trading knowledge.  Hence, I don't feel I have "wasted" my time in taking this "longer" path.  At least, I ended up with a strong and sound theoretical trading framework to rely on, truly know how to trade.  A person can search for a profitable EA and still end up clueless about how to trade.  That's the key difference between searching for profitable EA vs searching for trading knowledge. 

When I first started using FSB, my goal was to test out my trading concepts and refine it.  I believe, a strong trading theory, naturally and statistically will lead me to profitable EA.  Now I've achieved my initial goal, i.e. develop sound trading concepts, my next phase and aim of EA development is to develop EA for multiple trading strategies. 

I would like to work on using the Generator and together with different combinations of trading concepts that I've developed, to create different EA to suit different trading strategy/preference/style, eg scalper etc.  I'm rather excite with this prospect and potential.  Now this path would definitely be easier and shorter than my 1st journey.  Cos I merge both my trading skills and FSB's prowess in the development process.

I have great expectations on FSB's potential and capabilities.  I know as we all put our brains together and work towards giving contributions and suggestion and with Popov's ability to transform many of our request into reality and his passion to keep improving FSB, I am very certain the FSB we are going to have, down the road, will be a very different and much better trading tools than we ever could imagine it to be, since the beginning.

Re: Maximum Equity Drawdown

Hello Miroslav,
do you know already when you might find time to add "Maximum Equity Drawdown" in the Sorting possibilities (8) of the Repository Collections?

Re: Maximum Equity Drawdown

Ohhh, O forgot to fix that in this release.
I'll write it down to my ToDo now.

Thank you for your reminder.

Re: Maximum Equity Drawdown

Thank you, no problem, I still have many other things to test ;-)

Re: Maximum Equity Drawdown

Popov wrote:

Ohhh, O forgot to fix that in this release.
I'll write it down to my ToDo now.
Thank you for your reminder.

Do you have an idea already when the "Drawdown-sorting-criteria" will be available in FSB-Pro?
I think EA Studio has that already.

Re: Maximum Equity Drawdown

It would also be nice to have the equity drawdown in the filter on EA Studio (it's got the sort but I'm not even looking at a strategy if the dd is too large so high draw down strategies just clutter the list.)

Re: Maximum Equity Drawdown

Yes, I'll add Acceptance Criteria.

Re: Maximum Equity Drawdown

Hello Mirsoslav,
my Question is related to FSB-Pro!
The addition from Gizmo is o.k. but is cofusing my original request.
Can I please get an Idea when it will be added in FSB-Pro (eg. when is the next update expected)?
I teally would like to have that.

Re: Maximum Equity Drawdown

We will have an update before the end of August.
It appeared unsuspectingly difficult to add this option in the filtering criteria. I'll try to find a workaround.

Re: Maximum Equity Drawdown

Thank you for the fast answer.
I think we are talking about two different "nice to have"
FSB-Pro:
Filtering (Sort) criteria in the Generator setup is one point.
Sorting in the repository is the second point (which was my original request, maybe that is easier)
I would find it for example quite useful to run the generator in "system quality"-sort mode,
and then sort the results in the repository according drawdown
What you think about that?

21 (edited by yonkuro 2016-09-02 14:03:52)

Re: Maximum Equity Drawdown

Popov wrote:

We will have an update before the end of August.
It appeared unsuspectingly difficult to add this option in the filtering criteria. I'll try to find a workaround.

Hi Popov, how is the update development going? I hope everything is going well.

do or do not there is no try

Re: Maximum Equity Drawdown

>> Hi Popov, how is the update development going? I hope everything is going well.

Everything is fine. I just passed through all 103 integrated indicators;
- fixed calculation of "firstBar". Now it takes into account the Use previous bar value". Also the code is optimized to reduce the required bars for indicators that use other indicators.
- code cleanup. Legacy and deprecated notations updated. ex. iBar changed to bar.

There are literally thousand fixes. The new indicator versions will prevent crashes in the expert advisors related to the length of the data series.

Here is a GitHub link to dif of the last commit: https://github.com/PopovMP/FSB_Pro_Indi … d8df529837

Now, I have to make the corresponding cleanup and fixes in the MQL indicators.

Re: Maximum Equity Drawdown

Glad to hear that smile

do or do not there is no try