Discussion of article "Developing a self-adapting algorithm (Part I): Finding a basic pattern"

[MetaQuotes]

New article Developing a self-adapting algorithm (Part I): Finding a basic pattern has been published:

In the upcoming series of articles, I will demonstrate the development of self-adapting algorithms considering most market factors, as well as show how to systematize these situations, describe them in logic and take them into account in your trading activity. I will start with a very simple algorithm that will gradually acquire theory and evolve into a very complex project.

The EA features the ability to re-invest the earned funds. You need to use it. Until now, I showed the tests with conservative settings. But what if we set very aggressive settings and enable a lot increase? I am not a fan of high risks, but let's see what the algorithm is capable of. I will perform the test on GBPUSD from 2006.01.01 to 2020.11.25 in the "Every tick" mode. Of course, it is possible to test another symbol. The spread is reduced to 20. This is slightly above average. Figure 12 shows the backtest result for almost 15 years.

Figure 12. GBPUSD from 2006.01.01 to 2020.11.25, aggressive settings

As you may remember, the algorithm uses close prices. Therefore, this result is not a "test grail". In addition, the adequate spread of 20 is set. The algorithm's trading result on the real market usually coincides with the one obtained in the tester. I have never used it to trade with such aggressive settings. Besides, it is impossible to take into account real spreads in MetaTrader 4, so I will not argue that it would have passed this period as well in real trading.

Author: Maxim Romanov

[Zhongquan Jiang]

This is really a great article. Looking forward to the new article.

[Maxim Romanov]

Zhongquan Jiang:
This is really a great article. Looking forward to the new article.

Thank you, new articles have already been written and are in the process of being translated. They will be even more interesting.

[Michele Catanzaro]

Maxim Romanov:

Thank you, new articles have already been written and are in the process of being translated. They will be even more interesting.

Thanks very interesting, I'm waiting for the second one.

[Daniel Sastraamidjaja]

Interesting and great article! I wonder if this EA can work for a really long time?

[Gyula Szajlai]

Interesting concept and article, thank you!

I am working on a similar project, the main difference is that in the N-sized window (N is between 1-5 at first) I use the relation of HLOC values of consecutive candles and also the relation within all the bars in the window one by one. It seems this will make it too complex.

This concept is nice with its simpleness.

[George Muthomi Kimathi]

Great Read! Anticipating the next one!

[Eric Pedron]

Since you are using the second method, maybe it's good to change the topic because the algorithm given is not really self-adapting. The strategy looks interesting though.

[Maxim Romanov]

Eric Pedron:
Since you are using the second method, maybe it's good to change the topic because the algorithm given is not really self-adapting. The strategy looks interesting though.

yes, this algorithm is not self-adapting, it is the first step towards developing an idea.  There will be 4 articles in total and in the last two I will show you fully adaptive.

[Luis Leal]

You are searching for the efficiency, not for the equilibrium... For this work, are you sure that only the number of the candles is the unique place to see and if you will see too the average of the amplitudes between open and close could be more accurate and we can see other things...? You are treating the candles as a simple bet, like red and black in roulette, but the roulette is a closed circuit with almost 100% of efficiency and do not have opposite forces, only the 0s...
To have a self-adapting approach, we will be need an engine to see first the start point to counting and the duration time to count... The time and amplitudes of the the movements, averages, discards, the cycles... etc...

[Maxim Romanov]

Luis Leal #:
You are searching for the efficiency, not for the equilibrium... For this work, are you sure that only the number of the candles is the unique place to see and if you will see too the average of the amplitudes between open and close could be more accurate and we can see other things...? You are treating the candles as a simple bet, like red and black in roulette, but the roulette is a closed circuit with almost 100% of efficiency and do not have opposite forces, only the 0s...
To have a self-adapting approach, we will be need an engine to see first the start point to counting and the duration time to count... The time and amplitudes of the the movements, averages, discards, the cycles... etc...

I wrote 3 more articles on this topic, here are the links to them in order, read how the model developed and what I came to in my articles.

2 - https://www.mql5.com/en/articles/8767

3 - https://www.mql5.com/en/articles/8807

4 - https://www.mql5.com/en/articles/8859

I also have articles preceding this one, in which I bring them to the topic.

Naturally, I did not stop and continued to develop the theoretical model. Now I can already tell how the price series differ from a random walk, how to find these differences and what are the reasons for these differences. In the articles I have not described this yet, but read my next works, you may be interested.