Discussion of article "Creating a new trading strategy using a technology of resolving entries into indicators"

[MetaQuotes]

New article Creating a new trading strategy using a technology of resolving entries into indicators has been published:

The article suggests a technology helping everyone to create custom trading strategies by assembling an individual indicator set, as well as to develop custom market entry signals.

As we expected, the first testing stage did not show definite profitable areas. But, at the same time, one should pay attention to force index indicator. At M5 timeframe, the chart of trade profit dependency on indicator values slumps in the zero area. Importance of this observation is evidenced by the fact that this phenomenon appears at analytical indicator charts with all the parameters used for testing. For our template select parameters with the most apparent phenomenon character (maximal drawdown).

Let us zoom the chart analyzed. As you can see, the effect of this factor is available within the range from -0.01 to 0.01. The phenomenon observed is equally true both for buy trades and sell trades.

Author: Dmitriy Gizlyk

[FaraaMarz]

WellDone , Great Job!

I had this idea in mind for a few weeks , but I didn't have enough knowledge to Implement it !

Now you made my life easy and I go for further actions on the files

[Anthony Garot]

Interesting article.

[Dmitriy Gizlyk]

Thanks.

[Alexander]

Interesting article, thanks.

Important for the analysis are good historical price data. Where do you get the historical price data, from a broker?

[Transitus]

Sorry Bro, but honestly speaking havent seen a better way to create a perfect Data Mining Bias during my Phd. What you are doing is discriminative selection, in basics; assume that you have a cream have some berries on it which you dip your spoon to particular areas for catching berries and refuse to eat rest of the cream. You expect that next time you can do the same who else doesnt..? But what if Aunt Marry brings a new cup of cream with a different layout of berries on it ?? Believe me my friend.. Aunt Marry is much more linear and predictable then the nonlinear price action heteroscedasticity.

[gtyrozz]

Hi,

I really like this article. I do believe that backtesting is very important. But few factors keep spinning in my head.

I can't see how long positions were opened. What was initial capital (in those abstract units)? What position size was set?

It would add to the strategy some maturity as things that I have asked directly impact potential of any strategy- commissions. 

Each time you rolling over the day, you pay commission based on possition size. 

And profit of 1000 or 10000 may be huge or misserable depending on starting capital. 

I would appreaciate your reflection to this. Maybe I missed it somewhere in article.