Multi Currency EA - page 3

[Sergey Golubev]

Forum on trading, automated trading systems and testing trading strategies

Buying or Selling all 7 pairs

Sergey Golubev, 2024.10.14 16:05

Developing a multi-currency Expert Advisor (Part 12): Developing prop trading level risk manager

Throughout the entire series, we addressed the topic of risk control several times. The concepts of a normalized trading strategy were introduced, the parameters of which ensure that a drawdown level of 10% is achieved during the test period. However, normalizing trading strategy instances, as well as groups of trading strategies, in this way can only provide a given drawdown over a historical period. We cannot be sure that the specified drawdown level will be observed when starting a test of a normalized group of strategies on the forward period, or launching it on a trading account.

Recently, the topic of risk management was considered in the articles Risk manager for manual trading and Risk manager for algorithmic trading. In these articles, the author proposed a programmatic implementation that controls the compliance of various trading parameters with pre-set indicators. For example, if the set loss level for a day, week or month is exceeded, trading is suspended.

[Sergey Golubev]

Forum on trading, automated trading systems and testing trading strategies

Buying or Selling all 7 pairs

Sergey Golubev, 2024.11.16 07:31

Developing a multi-currency Expert Advisor (Part 13): Automating the second stage — selection into groups

Having been distracted a bit by the risk manager in the last article, let's get back to the main topic — test automation. In one of the previous articles, we outlined several stages that should be completed while optimizing and searching for the best parameters of the final EA. We have already implemented the first stage, at which we optimized the parameters of a single trading strategy instance. Its results were saved in the database.

The next stage is a selection of good groups of single instances of trading strategies that, when working together, will improve trading parameters — reduce drawdown, increase the linearity of the balance curve growth, and so on. We already looked at how to carry out this stage manually in the sixth part of the series.

[Sergey Golubev]

Forum on trading, automated trading systems and testing trading strategies

Buying or Selling all 7 pairs

Sergey Golubev, 2025.01.11 11:57

Developing a multi-currency Expert Advisor (Part 14): Adaptive volume change in risk manager

In one of the previous articles of the series, I touched on the topic of risk control and developed a risk manager class that implements basic functionality. It allows setting a maximum daily loss level and a maximum overall loss level, upon reaching which trading stops and all open positions are closed. If a daily loss was reached, trading was resumed the next day, and if an overall loss was reached, it was not resumed at all.

[Sergey Golubev]

Forum on trading, automated trading systems and testing trading strategies

Buying or Selling all 7 pairs

Sergey Golubev, 2025.01.17 07:15

Developing a multi-currency Expert Advisor (Part 15): Preparing EA for real trading

We have already achieved certain results in the previous articles, but there is still much to do. The end result we would like to see is a multi-currency EA that can be set to work on a real account or several real accounts with different brokers. So far, our efforts have been focused on achieving good trading results during testing, since without this it is impossible to achieve good trading with a developed EA on a real account. Now that we have more or less decent test results, we can look a little towards ensuring correct operation on a real account.

[Sergey Golubev]

Forum on trading, automated trading systems and testing trading strategies

Buying or Selling all 7 pairs

Sergey Golubev, 2025.01.31 18:16

Developing a multi-currency Expert Advisor (Part 16): Impact of different quote histories on test results

In the previous article, we started preparing the multi-currency EA for trading on a real account. As part of the preparation process, we added support for different names of trading instruments, automatic completion of trading when you want to change the settings of trading strategies, and correct resumption of the EA after restarting for various reasons.

[Sergey Golubev]

Forum on trading, automated trading systems and testing trading strategies

Buying or Selling all 7 pairs

Sergey Golubev, 2025.03.14 07:58

Developing a multi-currency Expert Advisor (Part 17): Further preparation for real trading

In one of the previous articles, we already turned our attention to the EA improvements necessary for working on real accounts. Until now, our efforts have been focused mainly on getting acceptable EA results in the strategy tester. Real trading requires much more preparations.

[Sergey Golubev]

Forum on trading, automated trading systems and testing trading strategies

Buying or Selling all 7 pairs

Sergey Golubev, 2025.04.15 15:48

Developing a multi-currency Expert Advisor (Part 18): Automating group selection considering forward period

We have implemented automation of the second stage of optimization using the forward period. Again, no clear advantages were identified. The task turned out to be much broader and required more time than we initially expected. In the process, many new questions arose that are still waiting for their turn.

We were able to see that if a forward period falls on an unsuccessful period of the EA's work, then we seem not to be able to use it to select good combinations of parameters.

[Conor Mcnamara]

Instead of making a new topic, I should ask in this one as my questions are on the same subject...

1. Can multi currency EAs be simulated in the strategy tester or does it only load the ticks for the symbol chosen?

2. Can the trade.mqh library be used to conveniently make a streamlined multi-currency EA? I have already tried something and I have no way of simulating the multi-currency part ...

The reason why I'm asking is because there are many articles (like those posted above) which is using a long winded way, or very custom way of doing things. I am wishing to use the trade library if it's possible, and it's better to ask to see if someone has already carried this out instead of me doing hours of live testing just to see something working with a custom indicator

[Conor Mcnamara]

Conor Mcnamara #:

Instead of making a new topic, I should ask in this one as my questions are on the same subject...

1. Can multi currency EAs be simulated in the strategy tester or does it only load the ticks for the symbol chosen?

2. Can the trade.mqh library be used to conveniently make a streamlined multi-currency EA? I have already tried something and I have no way of simulating the multi-currency part ...

The reason why I'm asking is because there are many articles (like those posted above) which is using a long winded way, or very custom way of doing things. I am wishing to use the trade library if it's possible, and it's better to ask to see if someone has already carried this out instead of me doing hours of live testing just to see something working with a custom indicator

Nevermind, got it working. I left "NULL" as the symbol name in the function call by mistake. Sooo it's very easy afterall to make a streamlined multicurrency expert advisor

[Sergey Golubev]

Forum on trading, automated trading systems and testing trading strategies

Buying or Selling all 7 pairs

Sergey Golubev, 2025.05.31 07:03

Developing a multi-currency Expert Advisor (Part 19): Creating stages implemented in Python

We accomplished the task and the hypothesis was confirmed. So, next looked at how we could improve the results of such automatic selection. It turned out that if we split the set of all single instances into a relatively small number of clusters and make sure that instances from the same cluster do not end up in it when selecting a group, then this will help not only improve the trading results of the final EA, but also reduce the time for the selection process itself. 

To perform clustering, we used a ready-made library scikit-learn for Python, or, more precisely, the implementation of the K-Means algorithm. This is not the only clustering algorithm, but considering other possible ones, comparing and choosing the best one, as applied to this problem, was beyond the acceptable limits. Therefore, essentially the first algorithm that came to hand was taken, and the results obtained using it turned out to be quite good.