Topic: Long, short, or both

It is a coincidence that we have a pair defined as EURUSD and not defined as USDEUR (there are only historical reasons why EURUSD and not USDEUR has been chosen). So what today is called long in the EURUSD pair would be, in the case USDEUR was the historical choice, called short. So, in principle, there should be theoretical symmetry in behavior logic for long and short dynamics of Forex data (I have to write “theoretical” because Forex market is often operated by people and people psychologically behave differently on long and short trades).

FSB seeks to find a winning structure in windings of a pair and to create a strategy which will work utilizing this structure. Logic created for long trades will be mirrored to create short trade logic. The compass of FSB is directed towards Financial North, where North will be the highest Net balance or System Quality Number or Sharp Ratio or some other chosen criterion. This search is done in the gigantic space of all possible strategies (by cooking statistically a stew on different indicators). The result of such a search is a strategy which I will call “double direction strategy”, as it is designed to work in both directions of trade.

Usually we will look at the results of this search process (a double direction strategy) in a glued form – we don’t separate results from long trades of a strategy and short trades of a strategy. But when I will do the separation and study the balance curve for the long trades and the balance curve for the short trades, I will often discover that those curves are very different. One direction (long or short) can make money, while the other will be just a drain. The sum of both will give positive results, but those results could be yet better if only one direction of trade has been chosen and the other has been removed.

A perfect double direction strategy should work equally well in both directions – long and short (equally only in a long run, and assuming that a pair is ruled by same logic in short and long trades). But mostly it is not the case. So I have been testing possibility to force FSB to search only in one direction of trade (by using “Long or Short” indicator).  I suspect that this way of generating a strategy should result in better strategies, because such strategy will be not cluttered (as a sum of two underlying behaviors).

Now I need your opinions on this, my fellow FSB travelers. Also opinion of our Capitan, Mr. Popov will be highly appreciated.

Let me end with my acronym AIRON? (Not yet known by everyone!) (Am I Right Or Not?)

P.S. Today there is a possibility to force FSB to produce strategies exclusively in one direction of trade by including indicator “Long or Short”. I think that this should be parameter controlled behavior, a criteria set in “Settings” or “Common” menu.

Re: Long, short, or both

There is in the program , in the 'Opening Logic' section, and indicator called 'Long or Short'

There are brokers that offer '' USDEUR''

The market will not allow us perfection in trading the same way in both directions, from what I have seen.

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....

Re: Long, short, or both

Mr. Blaiserboy, my long tirade could be cut down to the following: Is it better, in your opinion, to develop separate strategies for long and short trades? Or maybe it doesn’t matter for some reasons which I cannot see yet, and standard FSB approach is superior?

Re: Long, short, or both

If we talk about custom indicators, we can set them up on UsdEur, simple conversion of price data - 1/EurUsd. It is possible to convert the data file as well, but it can be used only in backtesting.

Certainly, if we take a weekly chart, in hindsight we can say on which direction we should have betted, if one believes in a "trend". Weekly trend is a big thing, like an ocean liner, rarely one can miss it completely. So there's a point to think about.

Re: Long, short, or both

My actual experience with creating FSB strategies is:

- Currency pairs - are symetrical and I develop exclusively "Long and Short" strategies.

- Indices, Commodities (lower timframes) - are more or less also symetrical, so "Long and Short" too.

- Indices, Commodities (M30 and higher) - I combine both methods.
"Long only" + "Short only" - basically bring higher returns; on the other hand there is a higher danger of creating over-optimized strategies this way.
"Long and Short" - if I succeed to create a good strategy, it is usually more robust, even if the total return is not so high.

Re: Long, short, or both

I think Bold has outlined my feelings, I use both approaches sometimes. Footon makes a good point about the weekly chart.

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....

Re: Long, short, or both

Regarding to symmetrical results or performance i donot think it can be achieved in Constance way

But :  by looking to The Strategy performance report i find a lot of data can be used and in this case

If Pop can add extra criteria that filter the  long group & short group at the same time for example

1 - Max short/long deviation '' in number of trades'' to guarantee that the trades is well distributed in both groups 

2 - % winners long group and % winners short So you can guarantee the strategy gives profits in longs and shorts even if one side is so bigger than the other in number of trades 

3- .....etc

Re: Long, short, or both

One proviso ....... the more we load the program............ the slower it will run when generating.

I am wondering if an 'add on' would serve to accomplish a number of the things we want for analysis....

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....