Advances in offline reinforcement learning methods show promise in tackling complex problems. However, handling optimism that arises during learning remains a challenge. This often results in overconfidence in strategies based on a training set data, which does not always encompass every potential environmental state and transition. This optimistic approach, especially in stochastic environments, could increase risks and lead to undesirable outcomes.

In the pursuit of a solution, autonomous driving research offers valuable insight. Notably, algorithms aim for risk reduction and user safety maximization, with minimal online training. One such approach, known as SeParated Latent Trajectory Transformer (SPLT-Transformer), presented in "Addressing Optimism Bias in Sequence Modeling for Reinforcement Learning" offers a unique perspective.

The SPLT-Transformer is a sequence generation mod...

Read more...

Understanding the intricacies of EA trading methods can provide a significant edge in the competitive field of market trading. The highlighted method, Breakout Range H1, comes with notable features.

Its primary functions include the application of a trailing stop, which supplements your trading strategy by securing earned profits. The method also integrates the use of a Moving Average (MA) filter to contribute to informed trading decisions. Additionally, it avoids auto lot sizing, relying instead on fixed lot preassignments.

Each trading move is supported by drawing lines of support and resistance, beneficial tools in identifying possible price action points. Furthermore, this approach advocates sequential completion of trades, maintaining a steady one-by-one transaction process for each trading period.

In the rapid-fire world of trading, using efficient, methodical strategies c...

Read more...

In this segment, the focus falls on the process of using machine learning in the creation of an Expert Advisor for trading and subsequent testing. The key objective of deploying machine learning algorithms is so computers can learn from data and improve with time. As part of the process, the decision tree model, favored for its simplicity and fundamental role in constructing advanced models, will be applied.

Our comprehensive guide discusses how to maneuver within the subsets of artificial intelligence, utilizing statistical modeling and machine learning to enable computers to perform tasks autonomously. Concentration will be on decision tree models, valued for their basic principles and elemental role in creating more complex data models.

Additionally, the guide deals with the integration of the Pandas library, preferred by data scientists for examining and processing data. Not to...

Read more...

The latest EA has been programmed to operate with multi-symbol capabilities, a crucial feature in contemporary trading systems. Configuring parameters according to individual trading strategies is straightforward, ensuring a broad array of options for investors.

One of the key functionalities includes a logical calculation of buying conditions. It focuses on price movements and compares whether they are higher or lower than the Middle_band. This operation works in conjunction with the volume analysis, where the previous volume is compared with the previous MA_volume.

Performance efficiency of the system is of paramount importance. In this development update, assurances have been given that comprehensive backtesting has been carried out. This task is part of an ongoing commitment to provide a secure and efficient system for every user.

The backtesting results compile as a testament...

Read more...

In the realm of Forex trading, it's essential to have access to the current local time of specific or all Forex markets. An impressively functional library exists that not only helps to get this information, but also aids in managing broker's server time and trading within particular sessions. This versatile library takes care of different time offsets and variations in daylight time changes, significantly easing the coding process.

Time zone is an unavoidable concept in this field, defined as the standard time correlated to a particular geographical location. This is determined by Coordinated Universal Time (UTC) which is a universally accepted standard. The understanding of UTC offset becomes crucial here, which varies based on whether the timezone is east or west of GMT.

To obtain the current time, the RefreshTime() method comes in handy by updating the current local time of the ...

Read more...

Implementing software features requires specificity. For efficient file storing, the specified path is designed for the 'Indicator' folder: C:\Users\HP\AppData\Roaming\MetaQuotes\Terminal\F762D69EEEA9B4430D7F17C82167C844\MQL5\Indicators\My Files\Donchian_Channel.

Remember, it is integral to add the requisite coding to the EA for proper operation. Careful and meticulous coding remains paramount to successful software development and operation. Following the guidelines set for file storing and code implementation will continue to uphold the high-quality work expected in this field.

Read more...

Introducing a new manual trading Expert Advisor (EA) aimed at those starting out in the world of trading. This specific EA is capable of being utilized in backtesting visual mode as well as during live trading. Notably, the EA includes the function to pause or adjust the review backtest's speed, offering users an improved environment for manual trading exercises.

The EA's features are robust and beginner-friendly;
-Lots Button: Enables lots quantity modification
-Buy Button: Places orders equal to the quantity of lots
-Sell Button: Places one short order for the selected lot size
-SL Modify TP Button: Modifies all orders relative to the current price of Stop Loss and Take Profit distance
-Additionally, buttons to manage and close Buy and Sell orders are provided

Furthermore, the EA comes with specifically coded parameters that assist users in implementing trading systems effectively...

Read more...

The segment that follows underlines paramount concepts and practices involved in refining the Simple Hedge Expert Advisor. By exploiting mathematics and employing a brute force approach, the process targets maximizing the performance of this trading strategy.

The article commences with an in-depth illustration of what optimization means in this context. Latterly, a combination of mathematical optimization and brute force approach is suggested to improve the outcomes drastically. The primary focus here is Mathematical Optimization and its implications in enhancing the Simple Hedge Strategy.

The text discusses the mid-term plan, starting with theory and calculations behind the optimization process, before shifting to a practical programming approach. The idea is to work on theoretical foundations first, then apply these to real code, thus ensuring an efficient transition from theory ...

Read more...

Invest further in programming prowess with this informative read. It starts with an emphasis on programming paradigms and the working essentials of procedural programming in MQL5 as well as functional programming. Ensuing detailed analysis sponsors a basic price action expert advisor employing the exponential moving average indicator and candlestick price data.

Delve into a rigorous examination of the object-oriented programming paradigm, subsequently adopting this knowledge to morph procedural code into object-oriented code, affording deeper comprehension of the distinctions between these programming paradigms.

The focus of this discourse is to elucidate the varying functionalities of different programming paradigms and their implementation in MQL5, with the secondary objective being the development of an easily usable price action expert advisor.

Here is also an in-depth persp...

Read more...

As a seasoned developer interested in making programming more understandable, it's been identified that breaking syntax limitations offers an approach to creating accessible and readable code. This openness might slow down the coding process, but the benefits become apparent in clarity and future comprehension.

Demonstrating this approach via the C_Mouse class, it’s revealed how this method broadens understanding beyond the standard programming language syntax. It’s encouraged for developers to create necessary definitions and adapt code to their personal cognitive style, promoting learning and innovation.

This approach simulates a translation process between different programming languages, establishing a wider understanding compared to conventional communication through standard symbols and terms. This concept expands your horizons as a developer, unlocking a realm of deeper under...

Read more...

Analyzing the principles of Charged System Search (CSS) for problem-solving in mathematical modeling and machine learning. Inspired by Coulomb's law in physics, CSS envisages possible solutions as charged particles that interact based on their charge and distance from each other. Through Newtonian mechanics, it computes each particle's position, aiming to strike a balance between exploration and exploitation of the algorithm.

The optimization power of metaheuristic algorithms plays a pivotal role here. Such algorithms stochastically search for near-optimal solutions to problems until a certain condition is fulfilled or a specific number of iterations is hit. This approach is considered as beneficial as it simplifies exploration of search spaces and decision making, without the need to devise new specialized heuristics.

The CSS algorithm maneuvers charged particles within this spher...

Read more...

Discover the intersection of algorithmic trading and artificial intelligence with the book "Neural networks for algorithmic trading with MQL5". Specifically designed for algorithmic traders, this illustrative resource offers insights into the integration of neural networks into MQL5 trading robots.

The book is comprehensive, spanning seven chapters that neutrally discuss everything from the basics of machine learning to advanced architectural solutions. Readers will have the opportunity to understand various types of neural networks, inclusive of Convolutional and Recurrent models, and explore the practicality of attention mechanisms.

Different methodologies for improving model convergence, such as Batch Normalization and Dropout, are elaborately dissected in this book. Practical guidance on the nuances of training neural networks along with embedding them into trading strategies ...

Read more...

The rather abstract field of Explainable Artificial Intelligence (XAI), reason out decisions made by machine learning models, poses distinct questions. As datasets burgeon and become intricate, it's challenging for data scientists to delineate algorithm behavior across all data horizons. XAI, in theory, empowers the building of trust in these models by explaining their functionality and validating their readiness for production deployment. Nevertheless, the explanations rendered should not be accepted blindly.

Machine learning comprehends relationships and interactions lurking in data, but determining these within the model calls for employing model explanation techniques. Insightful questions arise; which features are most informative? How does an output change with alterations in a feature? Which features warrant additional engineering? Answers, although valuable, confront formidab...

Read more...

Efficiency in trading strategies is instrumental; how one times the entries and exits can make a significant difference to returns. Presenting a concise snippet of code that enables time subtraction and could be employed to optimize trade timings. The key here is a tick counter, which executes the subtraction operation.

Bear in mind, the foundational date employed is 1970.01.01 00:00:00, a parameter that is notably versatile for multiple calculation requirements. This function may provide a considerable uplift in devising time-sensitive strategies, as the usage is incredibly straightforward. Implement it by adding it to the chart, followed by the visual prompt via a Comment with relevant information on screen.

Efficient code can be a game-changer in trading where precision meets profit. Learn and adapt, harness the potential of this programmable tool to strategize and win!

Read more...

Over a time of great evolution in software development, changes have been witnessed in the MetaQuotes team. Use of native WebRequest for MetaTrader 5 since 2014 signifies significant leaps. The transformation of tools such as IASS, PASS, and SAAS becoming mandatory in various software initiatives, makes the developer's and end user's jobs easier.

Significant changes have also been observed in database management. The once oddball approach of database access via REST API, has transitioned to the preferred method for myriad applications. The numerous providers of these services make them highly accessible and cost-effective.

The article provides an in-depth analysis on five alternatives to connecting a Postgres database to MetaTrader 5. Also, instructions on setting up a development environment, installing Postgres as a remote database, connecting to it and inserting/retrieving data ...

Read more...

If you've been looking for a comprehensive list of modified moving averages to use in your strategies, consider this compendium open to the IT community. This is not the original material - it has been brought here from the MQL5 forum, as it is otherwise unavailable to the public.

The algorithms include the Simple Moving Average (SMA), Exponential Moving Average (EMA), and Wilder Exponential Moving Average (Wilder), among a host of others. To further amplify its utility, price points used to construct moving averages range from open, close, high, low, median and weighted average, providing a versatile tool for trading predictions.

Tailor it to your preferences with the adjustable MA_Period defaulting at 14 for the averaging period of the moving average. MA_Shift equals 0 allowing you to shift the moving average by a certain number of periods, and ShowInColor can be set to ‘true’ en...

Read more...

In our constant pursuit towards achieving a 100% automated Expert Advisor (EA), minor revisions to the pre-existing segments of code are crucial due to potential interactions between the trader and the EA. Primarily, alterations in sections where data exchanges occur are fundamental. The factors to consider include robustness, reliability, stability, scalability, encapsulation, and efficiency of the system. The precise efficacy of the system depends extensively on having three core code directories: development, working, and stable. This structure aids refactoring and debugging, allowing programmers to pinpoint areas requiring modification or elimination.

Essential modifications have been made to the timing system and the order system. These changes provide greater stability and agility, respectively. Further steps delineate towards transforming the manual EA, ensuring proper automa...

Read more...

Discover the benefits of using current profit over pip distance for take profit in Expert Advisors (EAs). The EA NextBot code manages this effectively, particularly useful when juggling multiple open positions or using different EAs concurrently. One considerable advantage includes sidestepping potential pitfalls that a pip-based take profit might encounter, such as limiting profits due to broker slippage. For a comprehensive understanding of this, consider referencing the code of EA SwingBot.

The key to this approach is calculating the total number of open orders through a magic number, a unique identifier assigned to an order. A distinct code is then used to calculate the current profit by iterating through all open orders using a loop function. After each iteration, the order's profit adds to a variable thus accumulating the total profit. Additionally, minimum profit can be set ex...

Read more...

A detailed examination of the challenges faced when developing a replay/simulation system for trading is available. Key points include the importance of maintaining code reliability and stability, as well as the complexity of creating a replay/simulation system that mimics a real account's operation.

This assessment also identifies potential pitfalls of developing a replay/simulation system, such as dependencies in the main code and the lack of utility in creating a simulation if the same resources are not available when the real account is used.

Within this article, you'll find a breakdown of these complexities, including the necessary timer mechanisms for accurate bar trading simulation. It discusses setting a timer in a replay/simulation system, using the MetaTrader 5 platform as a case study.

Furthermore, it deals with the challenge of adapting time-focused codes and systems...

Read more...

Discussing the elaboration and working of the C_Mouse class, which applies the principles of object-oriented programming for smooth extension of functionalities. Untested parts can be implemented and tested in parallel, without affecting the main code. This reduces the potential for inadvertent bugs, offering advantages especially for large production projects. The article also emphasises the importance of using directory structures to revert back to earlier development stages when necessary.

There is further exploration of the integration of this with the C_Studies class, using pointers for extended functionality. Thus transforming minor projects into valuable additions to the final project, all while preserving the integrity of the class system. Analyzing two paths for system progress, with detailed code comprehension, this article provides essential insights for developers.

The...

Read more...

This update brings a notable implementation twist to the Commodity Channel Index, using exponential moving averages rather than the simple moving averages traditionally used by creator Donald Lambert. This change, practical and designed to enhance adaptability, demonstrates the evolving nature of programming methods.

This new method incorporates the use of conditional compilation, ensuring functionality on both MQL4 and MQL5. From a comparative standpoint, this strategy brings a significant advantage, enhancing compatibility and efficiency.

For those keen on keeping up with the latest CodeBase publications, rest assured that every piece of source code is readily available under the "Public Projects" tab in MetaEditor, listed under the name "FMIC". This easily accessible resource fosters a continuous learning environment to both seasoned and novel programmers alike. Keep pace wit...

Read more...