Discussing the article: "Developing a multi-currency Expert Advisor (Part 18): Automating group selection considering forward period"
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Check out the new article: Developing a multi-currency Expert Advisor (Part 18): Automating group selection considering forward period.
Let's continue to automate the steps we previously performed manually. This time we will return to the automation of the second stage, that is, the selection of the optimal group of single instances of trading strategies, supplementing it with the ability to take into account the results of instances in the forward period.
As always, let's first look at what we already have and what is missing to solve the problem. We can set the task of optimizing a trading strategy over any required time interval. The words "set a task" should be taken literally: to do this, we create the necessary entries in the tasks table of our database. Accordingly, we can first perform optimization on one time interval (for example, from 2018 to 2022 inclusive), and then on another interval (for example, for 2023).
But with this approach we cannot use the obtained results in the desired way. At each of the two time intervals, optimization will be performed independently, so there will be nothing to compare: the passes of the second optimization will not repeat the passes of the first one in terms of the input parameters' values. The same is true for the genetic optimization we use. It is clear that this is not true for full optimization, but we have never used it and most likely will not use it in the future due to the large number of combinations of optimized parameters.
Therefore, it will be necessary to use the launch of the optimization with the specified forward period. In this case, the tester will use the same combinations of inputs on the forward period as on the main one. But we have not tried running automated optimization with a forward period yet, and we do not know how these results will get into our database. Will we then be able to distinguish between runs in the main period and runs in the forward period? We should check this.
Author: Yuriy Bykov