Introduction to MetaTrader 5 and Programming with

MQL5

Create your 1st Investment Robot with MQL5 step by step from
ZERO.

All rights reserved. This e-book or any part thereof may not be reproduced or
used in any way without the express written permission of the author or
publisher, except for the use of short quotations in a review of the e-book.

First edition, 2018.

Author: Rafael F. V. C. Santos ([email protected])

Specialist in strategic risk management applied to the financial market. He

works with the development of automated investment strategies (Robots of
investment - Expert Advisor) using machine learning and spatial statistics.
Graduated in Chemical Engineering from the Federal University of
Pernambuco (UFPE) - Brazil. He holds a Masters and PhD in Civil
Engineering (UFPE) in the areas of characterization, modeling and statistical
simulation, applied to oil wells and reservoirs. He has several articles, with
the theme of applied statistics, published in national and international
magazines and congresses.

Copyright©2018 de Rafael F. V. C. Santos.

Summary
1. Introduction
1.1. Digital Age
1.2. Artificial Intelligence
1.3. High Frequency Trading - HFT
1.4. Computer Programming
1.5. What will we learn?
1.6. Why MetaTrader 5?
1.7. EA Strategy
1.8. Advantages of Using EA
1.9. E-book Summary
2. MetaTrader 5
2.1. Downloading and installing MetaTrader 5
2.2. Main elements of the MetaTrader 5
2.3. Navigator Field
2.4. Field Toolbox
2.5. Looking for Candels
2.6. Saving Templates
2.7. Drawing Objects
2.8. Transition between Graph Times
2.9. Adding Indicators
2.10. Data Window
3. MQL5 Community
4. MetaEditor MQL5
4.1. MetaEditor
4.2. Creating a new Project
4.3. OnInit()
4.4. OnDeinit()
4.5. OnTick()
4.6. Programming Accessories
4.6.1. Comments
4.7. EA Properties
 4.8. MetaTrader and MetaEditor File Types
4.9. Adding Libraries
5. Basic Programming Logic with MQL5
5.1. Types of Variables
5.2. Declaration of Variables
5.2.1. Integer Type
5.2.2. Type Double
5.2.3. Type String
5.2.4. Type bool
5.2.5. Type Datetime
5.3. Declaring Constants
5.4. Vector Variables: Arrays
5.5. For loop
5.6. Enum
5.7. Input Type Variables
5.8. Local and Global Variables
5.8.1. Local variables
5.8.2. Global Variables
5.9. Variables Predefined by MQL5
5.10. Math Operations
5.11. Logical and Conditional Relationships
5.12. Ternary Operator
5.13. Methods or Functions
5.14. Candles and Tick Variables
5.15. Functions Comment () and Alert ()
5.15.1. Comment()
5.15.2. Alert()
5.16. Adding MQL5 Indicators
6. Programming the EA
6.1. Strategy overview
6.1.1. Moving Averages
6.1.2. Relative Strength Index (RSI)
6.2. Creating EA
6.3. Declaration of Global Variables
 6.3.1. Variables for the User
6.3.2. Global Variables
6.4. OnInit () Function
6.5. Function OnDeinit()
6.6. Expert Advisor Functions
6.6.1. Error Handling
6.7. Function: OnTick()
7. Backtests
7.1. Backtest in MetaTrader 5
7.2. Analyzing the Backtest
7.3. Performance Charts
Chapter 1
1. Introduction
1.1. Digital Age

We are living in the digital age. Nowadays, information and knowledge travel
through the optical fibers. Everything is in constant development and
accelerated growth. This evolutionary dynamic is unprecedented in the
history of mankind.
Devices and objects are becoming increasingly intelligent and independent.
Nowadays, it is possible to find, for sale, smart phones and watches that
recharge the battery with even the temperature of the human body. Research
shows that by 2025, at least half of the US fleet will be traveling completely
autonomously. That's right, the profession of driver, taxi driver and many
others, are counting their days. And for the financial market this is no
different. Professional traders are being replaced by algorithms.
These are just small examples of the avalanche of modifications and benefits
brought in from the world shouting for change. We are changing the world
and the world forcing us to change into a virtuous circle of prosperity.
1.2. Artificial Intelligence

Embedded systems are gradually dominating our daily lives. People and
things are becoming more and more connected to the internet. A new
intelligence is created from this new cosmos. We are opening the doors to
Artificial Intelligence (AI). Practically everyone benefits (or will benefit) in
some way from the wonders brought by AI.
Many say that this new generation may bring us problems, but we must be
optimistic about the uncertainties, as we are on the way to an inevitable and
non-return path.
Several companies and investment funds are currently working heavily on the
use of artificial intelligence to make their investment decisions. From AI, we
can filter and group information and knowledge hidden in apparently
meaningless numeric data and connection.
In this race, it is advantageous to dominate subjects and knowledge related to
mathematics, statistics and computer programming.
1.3. High Frequency Trading - HFT

Investors and speculators in the financial market know that the main enemy
to make money in this world is precisely the emotional factor. Human beings
have desires, wants, greed, fears and fears that make the activity decide the
best time to buy and / or sell a real challenge.
When we are working with Investment Robots, we are discussing a strategy
that has been previously defined and structured so that it can be executed in a
timely manner. Fortunately, robots do not get tired, stress or have any
feelings that allow misunderstandings and / or biases when buying or selling.
Optical fibers allowed for a revolution in communication. In the investment
world, for example, every thousandth of a second can cost millions of dollars.
So-called High Frequency Traders (HFTs) operating on the world's top stock
exchanges makes decisions at infinitesimal speeds and thus are threatening
many professional brokers and traders.
The HFTs' decision to buy and / or sell is taken in milliseconds, up to ten to
fifty times faster than a simple wink of the human eye (lasting approximately
300 milliseconds). It is practically unfeasible and unfair the operational
comparison of any computational system of this category with a normal
human being.
Recent data (year 2018) show that more than 50% of US financial operations
are carried out by investment robots. And the results show that these robots
can increase financial returns by up to 15%. There are some stock exchanges
where more than 90% of operations are carried out by investment robots.
1.4. Computer Programming

Most people have the following objections when it comes to programming:

it's very difficult, needs a lot of practice, has to know a lot of math, I've never
been good with numbers. However, many of these objections are nothing
more than a misunderstanding of this programming activity, which is as
trivial as telling a brief story with beginning middle and end.
The interesting thing is that anyone able to write the steps of running a cake
recipe on a sheet of paper is able to create a computer program. To program
is to communicate with the outside, with the other, in order to conduct
actions. For our case, this other is the machine.
We have a structure of activities to be followed and we need, for a good
realization of it, an adequate sequence of order of execution. Programming is
to master a language the machine understands and know how to position our
deliberative needs in time and space, so that they are executed correctly.
Therefore, it is up to every curious investor and with the patience to learn
something new to look at in the world of programming. This is because
scheduling computers is a much simpler activity than many imagine.
1.5. What will we learn?

In this e-book we will learn how to create, step by step, an automated

investment strategy. We will create an Investment Robot (Expert Advisor -
EA) from scratch using the MetaTrader 5 and MQL5 programming language.
Thus, having a program written in a high-level language (MQL5) (those
closest to the understanding of humans, low-level languages are closer to the
understanding of machines) we will be able to leave the arduous and
exhausting work of monitoring purchase and sale of a financial asset on the
stock exchange for the robot.
We will spend a few hours to program a robot known as Expert Advisor and
spend the rest of the time monitoring the execution of the strategy in a much
quieter and safer way. That leaves more time for a contemplative and
pleasurable life.
1.6. Why MetaTrader 5?

MetaTrader 5 is a powerful platform for visualizing, operating and

scheduling investment strategies for various types of financial markets.
It was created in the year 2005 and continues to be improved and distributed
by MetaQuotes (https://www.metaquotes.net). We can use MetaTrader 5 to
operate in the stock markets, futures, options, currency pairs (FOREX)
among others.
MetaTrader 5 was chosen to build our first EA, as it is a 'free' tool, quite
powerful and complete from the point of view of the main functionalities that
a trader platform should have.
Inside MetaTrader 5 we have the option to create scripts, indicators and, of
course, investment robots using the MQL5 programming language. This
language is very similar to C ++, Java and C #. So anyone who already
knows any of them will feel extremely comfortable learning MQL5.
1.7. EA Strategy

We will develop from scratch a strategy of crossing two moving averages

(one fast and one slow) with the possibility of an input filter in operations
with the rather popular indicator RSI (relative strength index).
Our robot will be multi-strategy. It will be able to work, using as input
triggers in the operations, only the crossing of the moving averages, only the
RSI or both (crossing of averages plus RSI).
We will also learn how to make backtests of the created robot. Thus we will
be able to test various setups on historical data of various financial assets.
The book is full of guidelines, with screenshots, using the key parts of
MetaTrader 5 and MetaEditor.
We will learn the programming syntax in MetaEditor. We have several
features when programming with MQL5. It already presents a very
significant set of libraries to work with graphical manipulation, mathematical
operations, statistics, with price structures and control of the sending of
trading orders.
1.8. Advantages of Using EA

Let's go to the advantages behind using investment robots. They are many,
but we go to the main ones:
- Eliminates all emotional side;
- Test ready strategies with backtests;
- Simulate adverse market situations;
- EA are free from fatigue, stress, lose attention, or fail to execute the
strategy;
- We can optimize the strategy from a Robot;
- EA are infinitely disciplined and speedy compared to us;
- Statistical validations can be made. EA statistically validated are
winners and consistent in the medium and long term.
- EA has a single mindset that is to execute the scheduled strategy;
- EA operate from the first second of market opening to the last second
before closing, without truce and misconduct.
1.9. E-book Summary

So to achieve the goal of creating our first Expert Advisor, this e-book
organizes and simplifies in a didactic way the necessary steps with the
knowledge and usage information of MetaTrader 5 and its MQL5
programming language.
Here's a summary of what you will learn from reading this e-book:
- Learn how to install MetaTrader 5;
- Know the main benefits of the MetaTrader online portal;
- Understand the main features and differences between MetaTrader and
MQL5:
- Learn to add indicators and EA;
- Know MetaEditor and some of the main shortcuts to facilitate
programming;
- Learn the programming syntax of MQL5;
- Understand the main functions of the MQL5 development libraries.
- Create a strategy using trend indicators (moving averages) and consolidated
market indicators (Relative Strength Index - RSI);
- Program step-by-step a Multistrategy EA;
- Make backtests of the created EA;
- Know the backtest statistics;
- And much more.

Now we are moving towards the future of investments. Congratulations on

your decision to enter this grand and fascinating world of automated
investments.
Thank you for your confidence in purchasing this material. The future is here
and we will build it together.
Everyone a great read!
Important note[1].
Chapter 2
2. MetaTrader 5

MetaTrader is a platform developed to assist in the visualization of prices of

financial assets and execution of buy and sell orders on the stock exchange. It
is also possible to create scripts (algorithms for specific actions), indicators
and investment robots.
It is a free tool developed and distributed by MetaQuotes Softwares
(https://www.metaquotes.net/). Initially, it was developed thinking about the
Forex market. However, several brokerages have been allowing for the use in
the stock market, options and futures.

We need a company (investment broker) to provide routing pricing data for

the platform. Many investment brokers offer this service for free. Some even
offer a demo account where we can simulate buy and sell operations in the
financial market with dummy money. Also, in the demo account, we can put
our robots to test in real time before using them in a real account.
MetaTrader 4 was created in 2005 and since then has been undergoing major
transformation of improvements in its system and security features. It is
currently in its second major release called MetaTrader 5 which was released
in 2010.
We can use MetaTrader on both desktop and smartphones. It is available to
work on the Web, on Mac, Linux, Windows and Android systems. The
leading Internet browsers on the market support MetaTrader 5 online.

MetaTrader 5 is not only a tool for visualizing and routing orders (buy/sell)
of financial assets, but also an excellent platform for developing investment
strategies from scripts, indicators and robots (Expert Advisors). For this it
counts on the environment of development for the language MQL5
(MetaQuotes Language) that presents/displays great similarities with C ++,
Java and C #. Therefore, those who have had some knowledge in any of these
programming languages will feel comfortable with the MQL5 syntax.
In comparison to MetaTrader 4, MetaTrader 5 allows developing strategies
using the object-oriented programming paradigm (OOP).
2.1. Downloading and installing MetaTrader 5

In the world several brokerages offer free use of MetaTrader 5. All we need
is that we are a client of the investment broker. We need, after downloading
the platform, to request a release of a demo and / or real account so that we
can have access to the routing of the data of the financial assets.
The direct download of the platform can be done by the site:
https://www.metaquotes.net/en/metatrader5/.
As said before, several brokers around the world offer free routing, but many
for the Forex market. Some Forex brokers only need a valid phone number to
release a demo account. In this e-book we will use the systems with the prices
of financial assets present in the São Paulo (Brazil) BM&FBovespa stock
exchange.
However, to follow the studies in the book, you do not necessarily need to
have an account with a Brazilian investment broker.
The installation process is very simple and straightforward. After two clicks
on the file an installation wizard will guide us with the necessary steps.

Just go ahead until the end of the installation. At the end you will be asked
for an account with login and password. It is necessary for us to use the
pricing data for financial assets. In this e-book we will use a demo account
from XP Investimentos.
See below the open program requesting the login, password and type of
server (if real or demo).
2.2. Main elements of the MetaTrader 5

One of our main interests is to visualize asset prices from candles

(candlesticks). So, let's get to know the main visualization features of the
MetaTrader 5 platform.
In the menu: View -> Symbols, we have the option to choose the financial
assets that we wish to work.

We can choose between Bovespa and BMF Assets (continuous series). In the
assets of Bovespa we have the OPTIONS (calls and puts), the shares in A
VISTA and the INDICES (futures market) traded in the Brazilian market.
It is important to note that each brokerage will provide distinct data routing.
Some with more data and some with less. For example, if we were with some
brokerage that works with the Forex market, the options that would appear in
this window of Assets would be completely different.

The addition of a symbol that represents a company, which is being consulted

in the prices, can be executed in the Market Watch.
In this Market Observation field we can add the assets related to the
companies traded on the stock exchange in which the brokerage firm is
released to operate.
Still within the Market Watch field we have the Details tab that offers
additional information regarding the financial asset being evaluated. Below,
for example, we can see the information on this tab for Petrobras shares. In
the Trading tab we have several options where we can make purchases and
sales with just one click.
2.3. Navigator Field

The field Navigator is where it contains the information regarding the type
of account of the user if demo or real, as well as the indicators and robots of
investments available by the platform and those that were developed by
ourselves.
In this field, we can choose the type of indicator, script or EA and drag it
directly to the financial asset chart of interest.

Another way to add indicators or EA is from the menu Insert -> Indicators,
Insert -> Expert (Robots).
Let's add an indicator from the menu:

In the figure above, we can see that the indicators are separated by trend
indicators, oscillators, volumes, Bill Williams and Custom. Trend indicators,
as the name implies, are more apt to warn when the market is within some
well-defined directional behavior. On the other hand, the Oscillator indicators
are more adequate to warn of opportunities within a lateralized market, a
market in consolidation.
Bill Williams[2] is a very popular trader who has developed several important
and widely used indicators. Therefore, a specific section was created for the
indicators created by it.
It is interesting the possibility that we have to use indicators and native EA of
MetaTrader itself and indicators and EA created by us or even offered in the
Market tab of the field Toolbox.
2.4. Field Toolbox

In the Toolbox field we can find several importants informations regarding

the sale of the main indicators and investment robots developed for
MetaTrader platform. We can also download those provided for free (tab
Libraries). We can view the main news (tab News) in real time of trading as
market in auction and output of news relevant to the market.

In the Experts tab, we can view the main information logs regarding the
operation of our robots. In this place, we are informed about the sending and
execution of orders, problems with variables, errors of execution and
operations.
Now that we already know, in general, the main fields of MetaTrader 5, we
go to the main graphic field, which is the window of visualization of the price
candles.
2.5. Looking for Candels

The MetaTrader 5 platform allows us to customize the way we visualize

prices in different ways. For example, we can change the background of the
screen, the colors of the contours and fill of the high/low/doji candles, the
separation lines of periods, grids, among several other properties.
We can switch between three types of price visualization of financial assets.
These visual types are given through Bar Chart, Candlestick or Line
Chart.
2.6. Saving Templates

It is important to note that each time a new chart is requested to be opened in

MetaTrader it will appear on the screen with the default settings. However,
there is a way to save our screen preview customization from creating a
template.
For example, let's set up a template. For this we must make our
customizations. With a right click on the graphic field we can choose
Properties and make the appropriate color settings to our liking.

Let's consider the following changes:

Then right click on the graphic screen and go to Template -> Save
Templates. We chose a name and that's it.

So, just load the saved template, to a new graphic, so that the settings are
immediately applied. That way, we have a lot of flexibility to be able to
configure visualization templates for setups of different strategies.
2.7. Drawing Objects

You can draw several objects in price charts. We can add channels, support
lines and resistance, Fibonacci, up/down arrows etc.
In this way, it is possible to make a visual technical analysis very complete
with the main graphic objects present in the platform.

Right-click on the graphic you can see the list of objects and make different
types of edits like modify subtitles, colors etc.
It is worth spending some time familiarizing yourself with the use of graphics
and objects.
2.8. Transition between Graph Times

The transition between the graphical scales of candle formation observation is

quite simple and fast. Just right click on top of the chart, go to Timeframes
and choose the time of candles.

We can also change the candle time from the shortcut bar, as shown below:
2.9. Adding Indicators

The indicators, for many traders, are the main guiding factors for decision
making. These are the indicators that often alert us to enter and exit an
operation in timely situations. And when you talk about indicators,
MetaTrader does not leave you wanting.
Currently, the platform has more than 2,000 indicators available for free
download. This download can be done either by the site itself or by the
platform in the Toolbox field on the Code Base tab.
Let's, for example, add moving averages to the chart. We can do this either
through the menu Insert -> Indicators or directly clicking in the field
Navigator -> indicators, click and hold the chosen indicator and drag to the
graphic window.
As an example, let's insert a Moving Average indicator:

Now let's add two moving averages. Our averages should have periods of 72
and 32 with zero shift, exponential methods and applied to closing prices:
After clicking on ‘ok’, we have the immediate addition to the selected asset
graph.

Let's now add the Relative Strength Index (RSI). This indicator is interesting
as it is drawn in a separate window from the main candle price window (see
figures below).
Let's leave a period of 9 applied to the closing price:

The result is:

The RSI is an indicator used to warn about overbought and/or oversold.
Basically, we have two reference levels that we must choose for the
activation trigger of the overbought or oversold. We can choose the levels of
30 (oversold) and 70 (overbought).
For example, RSI values above 70 are indicating that the price is overbought,
that is, at any moment in a consolidated market, we can begin to see a drop in
prices. On the other hand, if the value of the IFR is below 30 we can say that
the price is oversold and a sudden high in prices is eminent.
These two indicators were presented, since they will be used to create our
investment robot. Going forward, we will use three strategies: one with the
crossing of moving averages, another with only the RSI and one with both
indicators.
2.10. Data Window

A very important field, which greatly facilitates the visualization of prices

and values of indicators, is the Data Window. To open it we can: go to the
View menu -> Data Window or use the keyboard shortcut Ctrl + D.

This action will open, for us, a new field: Data Window. Now we can
position the mouse cursor over any candle of the graph to obtain detailed
information of prices and indicators in the place chosen.

See the example below, where we place the mouse on the highlighted candle
in green and the Data Window gives us the detailed information:
Chapter 3
3. MQL5 Community

As discussed earlier, in Chapter 2, MQL5 is a programming language offered

within MetaTrader 5. Programmers often share codes and questions around
specific problems in a programming language.
It is very important to participate in communities of developers, because with
this our learning and the motivational side are enhanced. The sharing of
knowledge and doubts generates sustainability in our development and
growth as a programmer and user of the language.
It is always good to have a place to exchange knowledge, ask questions and
discuss project ideas. The more widely used and documented a programming
language, the better the learning curve for first-time sailors. Therefore, a
common environment where all are destined to the same end is quite valid.
In this sense, for the MQL5 language we are well covered, because we are
facing a mature tool, with a lot of developers and a very active and generous
community that offers articles and doubts with agility in the forums.
We can access the community from the site: https://www.mql5.com/en.
After reading this e-book and/or even now, the reader is invited to make an
account and start adding their knowledge and ask questions from the portal.
Soon after the registration we will have access to a very interesting set of
functionalities. For example, we can access the MetaTrader 5 web terminal,
documentation that is extensive and up-to-date, key information on the global
financial market, and other benefits.
Something very important for anyone who is starting with the MQL5
language are the open source code samples available in CodeBase. In this
location we will have access to codes for both the MQL4 and MQL5
languages. It is extremely valuable to spend time researching and studying
these codes.
We also have the section of Articles where tutorials with excellent didactics
are added. This is one of the indispensable fields for anyone with the pretense
of mastering the MQL5 language.
Anyone who wants to build an EA with a more robust and complex strategy
can hire MQL5 programmers from orders on the Freelance portal. Also, for
those who have begun to master the tool and want extra income, they can
accept projects and offer their robot production services, indicators and
investment scripts.

In the field Market we can buy indicators and EA offered by individuals and
specialized companies.
We have a very interesting field that is Investment Signals. Through it we
can follow the entry and exit of trader operations. We have the evolution of
yields, the level of reliability, maximum profit, among other information, for
each one of the traders who were willing to provide the signals of their
operations. Some signal services can be contracted (paid for) and automatic
execution can be synchronized with our accounts.
After finding a strategy and consistent setups and prove validity through
testing the demo account, an important step is to hire a virtual machine
service. This type of service will leave us more secure and protected with
regard to lack of energy and internet.

Oscillations in the energy and internet network are problems that can
compromise the proper functioning of our investment robots. An alternative
to solving this problem is by hiring a Virtual Private Server (VPS) virtual
machine. The MQL5 portal itself provides this kind of service.
There is the possibility of contracting the VPS virtual machine service by
other companies, such as Amazon AWS (https://aws.amazon.com/en/vpc/)
that even offers a free year of use.

We have the Forum field which is where we will probably spend a good part
of our learning journey answering and putting our doubts about the MQL5
codes. This is a good place to exchange experiences, make friendships and
partnerships.
Chapter 4
4. MetaEditor MQL5

The language MQL5 (MetaQuotes Language 5) has a large number of useful

functions to analyze the prices of financial assets. We need to know these
functions so that we do not recreate the wheel. But before it, is important to
know as much as possible of our language programming environment
(MetaEditor).
4.1. MetaEditor

MetaEditor is an Integrated Development Environment (IDE) for the MQL5

language that is included in MetaTrader 5. Let us now know the main
features of developments offered by MetaEditor.
To open MetaEditor we must click the highlighted icon in the image below or
press the F4 shortcut.

After clicking on the icon for MetaEditor we will have a development

environment with three main areas that will assist us in the development of
our codes.
Area 1: Navigator - where we will manage our code library archives and
libraries in development.

Area 2: Window where the codes are properly typed.

Area 3: Toolbox - here is where we will follow the development of our

compilation and debugging tests. We have 6 tabs (Errors, Search, Articles,
Code Base, Public Projects and Journal). Basically, for the creation of our
algorithms we will only use the Errors and Journal tabs.
4.2. Creating a new Project

Let's click the New icon and then choose from the MQL5 wizard the type of
project that we want to create.

We can notice that we have several design options. This Wizard even allows
us to create investment robots without programming knowledge. Of course in
this case we will be blind to most of the steps and we will hardly do anything
with confidence and due security.
As this ebook intends to provide the first steps with the MQL5 tool we will
only have the first option Expert Advisor (model). There are two[3] other e-
books from the same author that show you how to develop Indicator and
Scripts codes.
Let's first name our investment robot (Expert Advisor). We'll call it
EA_MA_Cros_RSI (Expert Advisor of Moving Moving Crosses with the
RSI Dindicator). This is just a temporary name for MetaEditor presentation.
After clicking Finish the MQL5 Wizard creates for us the basic body
required for the development of any investment robot. Therefore, we already
have a divided structure with the main methods for the operation of an EA.
The structure of an EA, for its proper functioning, will always contain at least
these three functions or methods: OnInit (), OnDeinit () and OnTick ().
4.3. OnInit()

The first function executed in an EA is OnInit (). It is triggered only when

we initialize the robot. We use this function to initialize pointers (Handles) of
indicators, variables, as well as make previous configurations and loads of
variables and templates.
4.4. OnDeinit()

The OnDeinit () function is called when the robot is removed from

MetaTrader. For example, we use this function to clean indicator handles,
variables, remove graphic and textual objects from the screen, and so on.
4.5. OnTick()

One of the most important is the OnTick() function because it is where a lot
of our programming logic is running. This function is one of the most
requested by our robot, since it is called every time a market operation
happens. At each new sell, buy, release and order cancellation this function is
called.
It is worth mentioning that this function is only called when we have EA
within trading hours. That's because only with the trading in progress is that
we have changes in the market.
If we are working on the development of a robot and we are not in the trading
hours we will not see any call of this function. Alternatively, we can use the
OnTimer() function to minimize these problems and to make tests and
debugs through it (using the Print() function). Later in Chapter 5 we will see
an example of using the OnTimer() function.
It is worth highlighting that there are several other predefined functions like
OnChartEvent(), OnTrader(), OnTester(), and others. However, we will only
study the main ones. Further details regarding these other functions can be
found in the MetaTrader 5 documentation.
4.6. Programming Accessories

We have some elements (functions) that help us in the readability and

organization of our robot codes.

4.6.1.Comments

Commenting on our algorithms is of the utmost importance for readability

and understanding of key steps.
For example, we have the comment element we can represent by // or / * * /

This type of structure (//) causes the compiler not to interpret the characters
they saw immediately. So we can make small reminders and explanatory
comments to make the code clear and readable for editing by other
developers and even us.
The second type of comment / * * / allows you to write entire paragraphs
between them.

We have some shortcuts to the comments. For example, if we want to

comment several lines we can make the selection of them and press (Ctrl +
~). To remove the comments, simply select and press (Ctrl + Ç).

The shortcut (Ctrl +.) Lets you create the following annotated string. This is
useful for separating large sections of code.

The shortcut (Ctrl +;) lets you create a small separation with the following
strings:
4.7. EA Properties

It is always good to inform details of robot development. Thus, directive

#property: serves to describe basic information about EA. In the example
below, where we have the property directive, we can observe information
regarding copyright, link directed to some website and version of the robot.

After pressing the Compile button:

We note that in the Toolbox field on the Errors tab the compiler did not raise
any problems.
So automatically in the MetaTrader Platform in the field Navigator ->
clicking on the more of Expert consultant we can find our EA compiled
(.exe5). You can now click, hold and drag to the candle chart and observe the
#property information present.

As we still have practically nothing programmed from the robot these

information in the Common tab above, are the only available so far. Then,
with each new modification made in the code with MetaEditor we can
observe the changes in the field Navigator in the part of experts.
4.8. MetaTrader and MetaEditor File Types

The file type generated with MetaEditor has the extension '.mql'. We can
verify this in the MetaEditor Navigator field where we have our robot edition
file called EA_MA_Cros_RSI.mql5.
After clicking on the button Compile a file with the extension
'EA_MA_Cros_RSI.exe5' is automatically created in the field Navigator ->
Expert Advisor of MetaTrader.
It is important to note the difference between MetaTrader and MetaEditor. In
MetaEditor we have files with extension '.mql5' and in MetaTrader we
have '.exe5' files.
4.9. Adding Libraries

We can add special libraries provided by MQL5. For this we just need to
know the name of the library with the following code structure:

The above example adds a set of classes and methods ready to generate
Fibonacci figures in the first include <ChartObjects/ChartObjectsFibo.mqh>,
and work with sending and managing sales orders in the second include
<Trade/Trade.mqh>.
Note that the extension of these include files is of type '.mqh'. MetaEditor
also allows you to create this type of file. From this we can create complex
algorithms that work with the graphic part and even visual animations.
We can research and study the main MQL5 libraries from the site:
In Chapter 5 we will know and study better MetaEditor functionality from the
study of the MQL5 syntax.
Chapter 5
5. Basic Programming Logic with MQL5

As mentioned before, the MQL5 language has great similarities with the C
++, Java, and C # languages. Therefore, anyone who has any knowledge of
these languages will have great ease in understanding the MQL5 operating
syntax.
Every algorithm, every recipe for cake, begins with the description of the
ingredients. So let's make an analogy and start studying the ingredients of our
EA, our cake.
Let's first understand the main types of variables present in MQL5. We also
need to know how this language declares and assigns values to variables.
5.1. Types of Variables

The variable is the basic unit of data storage in any programming language.
Almost every algorithm needs variables. In MQL5 we need to declare a
variable always with some type.
The main types of variables used to create robots are: Integer (int), floating
point - decimal values (float and double), alphanumeric characters (string),
logical (bool) and dates (datetime).
Unlike MQL5 there are some languages that have automatic typing with, for
example, Python.
Having to explicitly declare the variable type has practical security
advantages and makes it easy to find compilation errors. When we are
limiting the type of the variable, we protect it from possible arithmetic errors,
assignment errors, and programming logic errors.
5.2. Declaration of Variables

Every MQL5 variable must have a type and a name. The type of this variable
must always be specified before the name. Then we can assign values
corresponding to the variable type. Let's study the most used and that will be
part of the structure of our EA.

5.2.1.Integer Type

We have some peculiarities regarding the memory consumption required to

store a variable. In most EAs this is not a problem, but we will discuss these
issues in case you want to leave your robot optimized for the memory
consumption requirement of processing.
For example, we have the type variables char, short, int, uchar, ushort,
uint that serve for integer variables.

char - uses only 1 byte of memory. You can allow values in the range of -128
to 127.
short - uses 2 bytes of memory. Can support values in the range of -32,768 to
32,767.
int - uses 4 bytes of memory. You can allow values in the range of
-2,147,48,648 through 2,147,483,647.
uchar - uses 1 byte of memory. The range of values is 0 to 255.
ushort - uses 2 bytes of memory. The range of values is 0 to 65,535.
uint - uses 4 bytes of memory. The range of values is 0 to 4,294,967,295.

See the example below:

In this case we have a variable with the name ‘movingAveragePeriod, which
accepts integer values (int). We can see a value assignment equal to 24. If no
value is assigned to the variable the MQL5 immediately assigns zero.

5.2.2.Type Double

When we are interested in storing variables with decimal numbers the most
appropriate is to use the double type. The double type consumes 8 bytes of
memory. More economical, we have the float type that consumes half, 4
bytes.
The main difference between these two types (float and double) is that the
accuracy of significant digits is larger for the double type (15 significant
digits) while the float has only 7 significant digits. As the difference in
memory consumption for our robots will be irrelevant we will use only the
double type that is more accurate.
For example, the division, in the figure below, does not involve many
numeric digits. Therefore, the double or float type makes no difference. On
the other hand, if we had a problem where the result was some periodic
tithing multiplied by some number or other decimal, cumulative rounding
errors could interfere with the accuracy of the results.

To the present moment do not worry about not observing the results of
operations. We will do this later with the use of the Print() function. For the
time being, let's take a look at the structure of the statement of variables.
5.2.3.Type String

Textual character variables are typed with string. The MQL5 language allows
string concatenation in a very simplified way as shown below.

For a string if no value is assigned, an empty string (NULL) is placed.

5.2.4.Type bool

When the interest is to have a variable with boolean characteristics, that is,
that only allow true values (true) or false (false) we can use the type bool:

It should be noted that in the case the variable 'bool response' was declared
without initial value, so the value of (false) is immediately assigned. Only
then, see figure above, that we have a value change to true, because x is less
than y.

5.2.5.Type Datetime

This is one of the most widely used types, since asset price data are collected
over some time scale. We will always be able to make evaluations of times
and days for operations. Therefore, we should be well aware of the use of this
type of variable.
The date and time format for MQL5 is defined as follows: {D'yyyy.mm.dd
hh: mm: ss'} ie {year.mon.day hours: minutes: seconds}
The capitalized D letter placed at the beginning of the date causes the
variable to have the datetime constant.
5.3. Declaring Constants

We have a way to create constants from global declarations at the beginning

of our code from the #define directive.

Of course, the values of the constants can not be changed. If there is any
attempt to change the value, we will have a compilation error.
It is standard to consider the names of constants with upper-case characters.
5.4. Vector Variables: Arrays

An array is a variable that supports multiple values. We may think that an

array is something like a list of values with a certain type. Within this list we
can do iterations by traversing each of its values.

We will often use variable arrays to store mainly financial asset price
information and indicator variables.
Here is an example of a three-position array declaration for storing integer
numbers:

If we try to do:

This compilation error happens because we made an initial statement

(myArray [3]) to allocate only 3 memory locations and we are trying to add a
fourth position.
We can make our array a list of the type of variable we want. However, in
advance, MQL5 needs to know the size of the list of values we need for the
array. So putting some value in brackets soon after the variable name is
required.
Another way to declare arrays is to directly assign values to it. We should do
this in the following way:
5.5. For loop

We will now study a very important repetition loop used in the development
of investment robots. When we want the computer to perform repetitive
activities within certain predetermined limits, a practical way of informing
the machine is through the for loop.
When we type the word for in MetaEditor we will have a small set as shown
below:

When we press Tab, the editor will autocomplete the whole structure of the
for loop.

This structure means that we will iterate from the integer 0 (zero) to the
integer (total-1) with an iteration increment ‘i’ of a unit (i ++ is the same as i
= i + 1, as well as i-- is the same as i = i-1).
So if we want to make a scan inside the elements of an array we can do from
the for loop as follows:
5.6. Enum

The enum structure is a special type of integer where we can define, through
a list, assignment constants. With enum we can customize the types of
variables that best suit our problem.
For example, let's create an enum to represent the seasons of the year. We
know that there are only 4 seasons of the year: spring, summer, autumn and
winter.
When you type the word enum we have a small arrow pointing to the right,
as shown below:

Another Tab, we have autocomplete. Our enum should look like this:

We can now use ‘SEASONS’ as a type of variable to be declared. In many

situations this is very useful to make our code more intuitive and organized.

The value of output was 1 (one) because enum considers the ordering of
integers from zero. In our case we have spring (0), summer (1), fall (2),
winter (3). Interesting and quite useful the enum, right?
5.7. Input Type Variables

Input variables are the only ones that allow the user of the indicators or EA to
assign values to them. These variables are used to provide the user with the
possibility of changing the setups of the indicators, stop loss, take profit,
number of lots etc.
Input variables can be of any type including enum types. Let's take an
example to make it easier to understand.

After compiling the above code we will have in MetaTrader the following
possibility of setting the EA:
Therefore, the input directive allows us to dialogue with the user on the
Input tab of our EA.
5.8. Local and Global Variables

It is very important to know the distinction between local variables and

global variables. The confusion between these two types of variables leads to
many EA compilation and execution errors. Therefore, we must be very
careful and pay close attention to where we are making the declaration of
variables.

5.8.1.Local variables

Let's create a function to show an example of local variables. When variables

are declared within a function/methods they with their assigned values will
only be used within functions. So we have local variables because they are
declared within a method or function.

5.8.2.Global Variables

When we want a variable to be used by any function/method anywhere in our

algorithm, we must declare it globally as shown in the example below. Pay
close attention to this example as it is very important to understand the
difference between the variables.
Let's take another important example, now with the chaining of a structure
with the if-conditional Boolean. Carefully evaluate this example.
5.9. Variables Predefined by MQL5

The MQL5 language has several predefined routinely used variables for
strategy development. These variables have their own notation and are
preceded by an underscore (_).
The main variables most used and that we will apply to create our first EA
are:
_Symbol: represents the financial asset symbol present on the current chart
shown on the screen.
_Period: refers to the period, in minutes, of the current candle chart.
_Point: represents the size of points that characterize the financial asset.
_Digits: represents the number of decimal point digits of the current asset.
The use of these variables will become clearer when we are developing our
EA.
5.10. Math Operations

We have already seen several mathematical operations. Let's study some

more of them. The MQL5 language provides us with a very intuitive and
simple structure for working with mathematical operations.

The language itself already comes with a set of mathematical and statistical
functions to work connected with the most diverse operations. We have the
functions with radical MathXXX quite useful:
5.11. Logical and Conditional Relationships

In the structure of our algorithms we commonly use logical relations that

return values of true or false. The main relationships we will be working on
are represented as follows by MQL5:

See the use of the if conditional plus the logical connectives && (AND) and
|| (OR):

Let's now go to the conditional structure of the else type. The else is activated
if the ‘if’ conditional is not true. Therefore, if the ‘if’ is false automatically
the else becomes true and everything inside the keys of this structure is
executed.
The else if structure is very useful in many cases. When the chain of
conditions is extensive we can use this structure. If no else if is true, else
executes automatically.
5.12. Ternary Operator

The ternary operator exists in several high-level programming languages and

in MQL5 it is no different. This is a very useful operator to simplify
conditional structures that could use if and else.
The ternary operator has the following operating structure:
(conditional)? (alternative 1 if true): (alternative 2 if false)

Let's take an example:

In the place where we use the conditional variable we could do any structure
using the logical evaluators (> =, <=, &&, ||).
5.13. Methods or Functions

When we are working with object-oriented programming (OOP) we make

constant use of classes. Classes are the master plans of the objects
architecture. Do not worry if this seems complicated because we will not use
OOP.
Among the classes we usually find methods and attributes. The term function
is most commonly used when we are working with procedural programming.
The MQL5 language allows us to program both in the object-oriented and
procedural paradigm. This book is about procedural programming, so we will
talk about methods and functions as if they were the same thing.
But what are methods or functions? They are code blocks that allow you to
perform a specific task. Usually the functions have input and return
parameters. In other words, the functions process the parameter entries
according to a well-defined logic and returns a result.

Here is an example to further clarify all this:

In a function we can have several types of returns and input parameters. The
processing performed depends on the logic of the problem.
Let's take another example where we have functions used to calculate the
sum, multiplication, and division of two numbers:
Functions that do not return values have a void type declaration. However, if
the function has some kind of return we are required to return something
relative to the return type.
5.14. Candles and Tick Variables

The MQL5 language has a set of libraries prepared to work with the
acquisition and manipulation of candle prices. We have the MqlRates
structure that stores the price information of opening, closing, maximum,
minimum, volume, spread and time.
Tick information refers to those collected from the book of offers such as ask
(bid), bid (sale), last business etc. This information can be accessed from the
MqlTick structure.
Let's see how to declare a candle variable and tick. We must beware, because
we must first declare the variable after doing the proper loading of the data in
the variable declared from other specific functions.
In order to load, we must specify the financial asset (which can be the current
one of the graph using _Symbol) and the period of the candles (to use the
current chart we put _Period).
For an example, we will create a new Expert Advisor with the name of
Study_Var (study of variables). In the wizard now we will add the OnTimer
option.
We decided to add the OnTimer() function, because with it we can work and
view our debug codes outside the trading hours. As we saw in Chapter 4, the
OnTick() function is called every time a new trading operation is performed,
and as we are probably making the codes outside trading hours no call to this
function will be made.
So we'll use the OnTimer() function to make timed calls and be able to see
the development of our codes. Let's ask it to be called every 2 seconds from
the EventSetTimer(2) function (see example below).

The CopyRates() and SymbolInfoTick() functions have been completed

following the following structures:
We can better understand the structures and main functions of the MQL5
library by using the F1 key. To do this, just place the mouse cursor on the
structure of interest and press F1. For example, let's do this for MqlRates:

From the F1 key we have access to a very complete documentation of the

MQL5 language. So we should always use it when a doubt arises.
As already discussed, we are not always developing EA in the trading hours
and even then we need to make some debugs in the code, so we are using
OnTimer().
However, within the OnTimer() function we will add the following code:
Let's compile. Now let's go to MetaTrader and drag EA 'Study_Var.exe5' to
the chart. Then in the Toolbox (if it is not showing go to the display menu and
ask to open) and enable the Experts tab.
Once the robot has been added to the graph, we can notice that on this tab
every 2 seconds we have an update of the OnTimer() function and what is
inside it is executed. In our case we see the prices of open, max, close and
min.

We are taking the prices of the candle in position '0', that is, collecting the
values of the most current candle. The candle can represent any time chart,
what really matters is your position. We can choose any desired position, all
we need is to understand the ordering of the vector. Below is an illustration to
facilitate this understanding (it is worth noting that this illustration is valid
only after using the ArraySetAsSeries() that orders the vector as shown).
Our candle vector has only 10 positions, because we use the function
CopyRates(_Symbo, _Period, 0,10, candles).

We can make our code more easily able to choose the position of the candles
and, in addition, add tick information, see below:

Let's now test the case of the variable (pos = 10). We chose for candles a
vector with only 10 price positions, but we are trying to access the position of
number 11 (remembering that the count starts with zero). So the code will
compile normally, however when executed in MetaTrader we will get the
following error message:
We have an array out of range error. MetaTrader also informs the line of code
in the .mql file where we find the violation (line 59). We must be alert
because this type of error is quite frequent in our robot creations. So be very
careful with the sizes and tracking positions of the arrays.
5.15. Functions Comment () and Alert ()

By using some useful functions we can create warning elements and alerts to
provide greater security of use and better understanding of what is happening
with our EAs.
5.15.1. Comment()

Comments are always welcome. Let's take an illustrative example:

We are using the DoubleToString() function to make the change from

double to string. If we are not careful to use this function we will not have a
compilation error, but in the Errors tab of MetaEditor Toolbox we will
observe the following composition of warnings:

5.15.2. Alert()

The Alert() function displays an alert box in the center of the screen with the
requested information. Those who are interested can search for audible alerts,
because it is also possible to add them to our EAs.
5.16. Adding MQL5 Indicators

Let us now turn to one of the most common and important practices when the
goal is to plan an investment strategy. It is very likely that your investment
strategy will make use of some indicator, so let's learn how to call indicators
from the MQL5 language.
Initially, we need to declare the variables for our indicator. How we will
declare variables will depend greatly on the type and functioning of the
indicator.
There are indicators that provide only one output of information, such as a
moving average. From the moving average we have only one output of
information which is the value of the moving average calculated for a specific
point in the graph. In this case we say that the moving average has only one
memory buffer. Every indicator needs at least one Handle and it must be an
integer type.

The declaration of the moving average indicator is given by a function of

MQL5 (iMA) with a F1 in this function we obtain:
In our case we are using a moving average period of 7 with the simple
smoothing type (MODE_SMA) applied to the closing price
(PRICE_CLOSE). It is good to point out, once again, that each indicator has
its own structure of parameters necessary for its proper functioning.
It is always good to put codes to evaluate possible variable loading errors. In
our case we added an Alert() if we have problems with the variable
'mm_Handle'. We did this because the correct functioning of this variable is
extremely important to our robot strategy. However, the practice of
evaluating the loading of variables is always welcome.
The same procedure can be done for the iRSI() indicator.
We made the initial statement and addition of the indicator on the chart with
the function ChartIndicatorAdd(0,0,mm_Handle). Now we should point
'mm_Handle' to the vector of ‘mm_Buffer[]’ which is where we will access
the values of our moving average.
To do this within the OnTick() function we must call the CopyBuffer()
function.
We can understand the structure of CopyBuffer() as follows:

Now let's add the RSI oscillator indicator. This indicator also has only a
single Buffer. If you are in doubt regarding the number of buffers of each
indicator simply go to the documentation from the F1 key and look for
#property indicator_buffers:

It follows the code with the addition of the two indicators (moving average
and RSI). Note that the (iRSI ()) function of the RSI indicator has input
parameters that are very different from the iMA () function:
Finally, we are fully equipped with the minimum information necessary to
start programming our investment robot. Let's go to the next chapter!
Chapter 6
6. Programming the EA

Now that we know the main functionality of MetaTrader, we have already

understood the basics of the features of the MQL5 language and the use of
MetaEditor we are minimally prepared to start developing our investment
robot (Expert Advisor - EA).
Before starting to program, let us know and understand the details of our EA
strategy.
6.1. Strategy overview

Our EA will have two moving averages and the RSI indicator. The crossing
of two Moving Averages works well when the market is in a trend. On the
other hand, in parallel markets or in consolidation, the RSI indicator performs
better.
There are several indicators that offer indicative when the market is in trend
as the ADX indicator. We decided to choose the RSI for reasons of making
an alternative between the possibility of choosing the market in the trend
and/or lateralized market.
For the buy or sell operations, our EA will have the possibility to choose only
one of the indicators (crossing of averages or RSI) or both at the same time.
Thus, we will have an EA that will track the input triggers based on two
indicators: moving averages and RSI.
Let us understand how the input triggers in operations will be given for each
of these indicators.

6.1.1.Moving Averages

First we go to the trigger for the crossing of moving averages. We have two
moving averages and the type, whether exponential or arithmetic, of these
averages can be changed as we will see next.
One of the moving averages will be called the fast average and the other
slow average. The fast average is the one that has the shortest period, on the
other hand the slow average should always have the longest period.
We give these names to averages because fast averages are those with a
shorter period of time and are more adherent to price, so they move faster
than those of longer periods where they have slower directional behavior.
Let's look at an illustrative diagram of how these triggers work in trading:
6.1.2.Relative Strength Index (RSI)

O indicador RSI (Relative Strength Index) no MetaTrader se refere a um

Oscilador, ou seja, é útil para indicar posições de sobrecompra e sobrevenda
em mercado lateralizado (consolidado).
We will not go into detail about the calculations behind the RSI because it is
not the goal at the moment. But let's understand the kittens buy and sell this
indicator
The figure below shows the RSI and two main levels (solid lines in red) at 30
and 70. Each of these lines represents the levels to allow overbought (line
70) and oversold (line 30). In other words, in a consolidated market, when
the RSI is equal to or greater than line 70, it can be a reversal warning,
warning an overbought, that is, it may be a good time to make a sale. On the
other hand, if the RSI, once again in a consolidated market, is less than or
equal to line 30, we are in an oversold region indicating a good time to buy.
In our EA we will leave those levels of 70 (overbought) and 30 (oversold)
possible to be changed by the user. So you can choose more loose or
restrictive levels for the kittens to buy and sell.
Let's now set an example where the trigger levels to buy are at line 30 and the
trigger to sell is at line 70:

Let us, therefore, to a summary of the strategy with the elements that we will
need to program for our EA:

1 - We must add two moving averages (fast and slow) both with the
possibility of choosing the period, the average application price and the
average method used.

2 - We must add the RSI indicator with possibility of period choice, applied
price, and oversold and oversold levels.

3 - The user can choose among the options to enter the traders if they come
from the moving averages plus the RSI, only moving averages or only the
RSI.
Now that we know the operation to buy and sell for our two indicators
(crossing of averages and RSI) we are ready to start developing the algorithm
with the presented strategy.
6.2. Creating EA

Let's open MetaEditor and ask the wizard to create a new Expert Advisor.
Let's give it the following name: MM_CROS_RSI.mql5.
If you want to obtain the complete source code of the robot you can access
the following page in github https://github.com/rafaelfvcs or send an email to
[email protected]. However, it is strongly recommended that you enter
the codes presented to gradually memorize the use of MetaEditor functions
and features.
Here we have the skeleton of our EA:
6.3. Declaration of Global Variables

Let's start by declaring our global variables. We will separate those variables
that the user will have access to modify and those that will be used for the
structure of the EA algorithm.

6.3.1.Variables for the User

We will provide the following options (input parameters): profit targets, stop
loss, number of traded lots, chart time, average periods, RSI period,
application of methods in prices (opening, closing, max, min) for both the
averages and the RSI, possibility of choice in the strategy of entry triggers in
operations (Moving averages plus RSI, only crosses, RSI only) and time limit
to close the open operations.
Below is the code of all this:

For the 'strategy' variable it was chosen to create an enum (this should be
declared at the top of the algorithm above the Input variables):

6.3.2.Global Variables

We must now declare the global variables that will be used by the functions
of our algorithm. See below:
6.4. OnInit () Function

In the OnInit() function we will add the initialization variables to the

indicators and add them to the chart with the ChartIndicatorAdd() function.
Follow the code:

Note that ChartIndicatorAdd(0,0,mm_fast_Handle) and

ChartIndicatorAdd(0,0, mm_slow_Handle) has the same graphical addition
reference (0 zero), which means that both indicators will be on the main
screen where the candles . However, ChartIndicatorAdd(0,1,ifr_Handle)
terms the reference (1 one) ie the bookmark will be added to another chart
box.
6.5. Function OnDeinit()

When the EA is removed we need to remove the indicators and if necessary

clear variables and parameters. The OnDeinit() function is called when the
robot is removed. Thus, within this function we can write our disinitialization
activities of the parameters:

As we write these codes it is advisable to compile and test EA in MetaTrader.

So if we drag the file: MM_CROS_RSI.exe5 in MetaTrader for some
financial asset we will have the following:
Clicking on ok will add the indicators to the chart and the EA will appear in
the upper right corner or not if the automated trading button is on.
See the example below:
6.6. Expert Advisor Functions

Let's create some useful and important functions for the operation of our EA.
We can use the shortcut (Ctrl +.) To create sections for a set of functions that
we will create for our EA.

First, we will need a function that informs the appearance of a new candle
regardless of the graphical time. Still, we do not have a native function in
MQL5 that allows us to accurately assess changing candle and so we must
create it.
The function should return true if a new candle appears on the chart. Here is
the code used in the tutorials in the MetaTrader article portal:
Let's do a function to help you visualize the entry points of our strategies.
This function basically will construct a graphical object (vertical line). Every
time our robot presents logical conditions (the trigger is activated) a vertical
line will be drawn on the graph with a name and color referring to the type of
warning.

The function drawsLineVertical() has four input variables. MQL5 allows

the last input variables of a function to already come with a default value. See
the case of the variable 'color = clrAliceBlue'. This means that when the user
does not set a specific color the 'clrAliceBlue' will automatically be assigned
to the variable. Therefore, the user has no obligation to assign values to this
type of variables. However it is mandatory to enter values for the other
variables.
Now let's create one of the most important functions for our EA. They are the
functions of sending orders.
We need to send orders to buy or sell every time a trigger is started. In MQL5
we can do this type of sending in different ways. We can use a specific class
called CTrade to work with the handling and sending of orders. However,
here we choose to use the requisite assembly structures (MqlTradeRequest)
and send response (MqlTradeResult).
MqlTradeRequest is used to perform all trade requests for a trade operation.
It contains all the fields necessary for performing a trade deal.

MqlTradeResult is used for any trade operation. The variable declared to be

of this type will be able to access the trade request results.

From these structures we can delimit the type of order that we are interested
in, specifying details of it. The request and response variables must be used in
an order send function called OrderSend().
Below is a function created to send an order to buy to market with type of fill
FOK (fill or kill - or fills the lot size of the order or does not execute). This
order, executed to market, already enters with the stop loss and target of gains
(take profit) defined.
In every market there are offers from the best buyers and best sellers.
Therefore, we need to access the book of offers to know these prices. As we
have seen, in Chapter 5, we can access book information from the MqlTick
structure. So we created a variable called 'tick' to store this information. With
it we can access the best seller 'tick.bid' and the best buyer 'tick.ask'.
We have the bid = (sell, offer) and ask = (buy, demand).

The type of action to be chosen is the first requisition that we should assign:

We have several types of orders. We can choose from the structure type:
The NormalizeDouble() function is used for rounding variables. We need to
use it because some financial assets have the number of different price digits
of others. We have assets that are quantified by points (as is the case of index
and dollar futures contracts), others are quantified by price and even pips
(Forex market). In Chapter 5 we discussed the use of _Point and _Digits.
The values of the SL and TK variables should be chosen by the user in the
form of points. Notice that for the stop loss price we did:
NormalizeDouble(tick.ask - SL * _Point, _Digits) as we are buying and our
stop loss point needs to be below the (ask). Already the target of the take
(take profit) naturally for a purchase should be above TK ask points.
It is with the request request.type_filling = ORDER_FILLING_FOK' that we
choose the order fulfillment type. We basically have three types of order
fulfillment:
Let's see the function to sell:

In this case, we have the stop loss level above the bid price
NormalizeDouble(tick.bid + SL * _Point, _Digits) and the gain target
naturally below NormalizeDouble (tick.bid - TK * _Point, _Digits).
We also need to create two functions to close open orders if arrive at the time
limit chosen by the user. That's because our robot is daytrader, we do not
want to sleep positioned. Remembering that to guarantee the closing of open
orders we have the variable 'Limit Closing Closing Position' chosen by the
user.
So to close a sale we must buy and to close a purchase we must sell.

We can note that for the CloseBuy() function the request.type_filling =

ORDER_FILLING_RETURN', this is because we want to close the order
independent of the best buy or sell. As we know this type of fill
(ORDER_FILLING_RETURN) executes the order until it reaches the limit
of the number of lots.
6.6.1.Error Handling

It is always good to treat the returns of the orders sent. This is because we can
minimize problems and evaluate execution errors. In our functions we did
this from answer.retcode which returns a code (with 5 digits) coming from
the brokerage server. Below is a window showing details of these codes. You
can get to this window by pressing the F1 key on code number 10008 or
10009.
6.7. Function: OnTick()

The goal now is to create the whole strategy logic for our EA. We already
have the necessary functions ready for its operation. It is within the OnTick()
function, called each new business carried out on the stock exchange, that we
will write the strategic body.
First, we must feed our candle, tick, and indicator variables with data. Also,
let's not forget to sort them appropriately with the ArraySetAsSeries()
function:

After this we can make use of these variables to build the logic of the buy and
sell action of our EA. Follow the conditions of crossings:
In the MQL5 language it is:

We have to create a logic to use the enum (STRATEGY_IN). We need the

sales and purchase triggers to be oriented with the options of: ONLY_MM
(mobile averages only), ONLY_RSI (RSI only) and MA_AND_RSI (Moving
averages plus RSI) described in this enum. Here's a possible solution to this
problem:
Now we can organize our functions so EA can make buying and selling
decisions. It is important to know when a new bar (candle) is created so that
the algorithm can track, without much computational effort, a possible
crossing of averages and RSI values in the desired region. Let's draw a
vertical line when the buy/sell triggers are activated.
The 'PositionSelect(_Symbol)' function returns a boolean. If there is any
position in progress it returns true, otherwise false. So we just want to get
into an operation if the EA has no open position.
Finally, to end our long-awaited EA we need a code to close some open
position at the time limit, i.e., that position that did not reach the winning
target or the stop loss.

The PositionGetInteger(POSITION_TYPE) function returns the open

position type. If it is a buy position we have POSITION_TYPE_BUY in the
case of a sale we have POSITION_TYPE_SELL.
Once again if you want to have access to the complete code of this EA can be
downloaded at: https://github.com/rafaelfvcs. Or if you prefer, you can send
an email to [email protected] and I'll be happy to send it to you!
Let's now do backtests with our newest multi-strategy investment EA.
Chapter 7
7. Backtests

Our main interest with Expert Advisor (EA) is to have a winning strategy
over the long term. There is that famous phrase: that nobody in the financial
market has a crystal ball. It is no use trying to predict the future of the market,
because unfortunately, it is impossible until the present moment.
We need to be realistic to accept that our strategy will have ups and downs
regarding financial returns. The 100% winning strategy is a utopia. However,
it is important to know how good our strategy is. There is nothing better than
using past data to gauge how it performed under various market conditions.
So testing our strategy at past time periods is called a backtest.
Fortunately, when it comes to backtesting MetaTrader 5 provides a collection
of simple and useful features to extract as much information as possible from
the performance of our EA.
Now that we already have an EA (MM_CROS_RSI.exe5) multi-strategy
programmed and working properly, let's do the backtests.
7.1. Backtest in MetaTrader 5

Let's open MetaTrader and go to the field Navigator -> Expert Advisor
consultant and look for our newest EA. Once found we should right click on
it and then on Test.

This will give us a new field in MetaTrader called Strategy Tester, in the
settings tab we have the following panel of possibilities:

With these features we will be able to choose the financial asset for the
backtest:

The graphical time:

We have the option to choose how the formation of our candles are
simulated by MetaTrader.

For this way of generating prices in candles, let’s go to the main options most
used:
Every tick or Every tick based on real ticks: MetaTrader tries to simulate
real price ticks in the formation of candles. This option leaves the backtest
more accurate, but requires more computational resources once inside a
candle the MetaTrader tries to reproduce the formation of real ticks. It is
worth noting that ticks are not real are just a simulation (a possible
representation of reality). This article clarifies the functioning of this
simulation:https://www.mql5.com/en/articles/75.
1 minute OHLC: In this type of candle formation MetaTrader generates only
4 price values: O - open, H - High, L - Low, C - Close. This option requires
much fewer computational resources compared to the tick option.
Open price only: This is the least costly option computationally, because it
only uses a price to represent a candle.
In our backtests we will always leave the option Forward marked No (see
the figure below). This option is useful when we are doing strategy
optimizations and want to use optimization methods such as Walk Forward
Analysis (WFA). This type of study is beyond the scope of this book. If you
want to know more about it, you can buy the other e-book (How to Properly
Optimize Investment Robots in MetaTrader 5[4]).

The type of execution with or without Delay is interesting so we can

simulate the conditions in which the servers are and the speed of our internet.
Latency refers to the time required for routing orders to be properly executed.
We have the possibility to choose the financial from the initial deposit as well
as the possibility to leverage investments (in the case of the figure below 1: 1
- without leverage).

For our backtests the Optimization field should be disabled:

We should choose the data period for the backtest.

And check the option to view the backtest:

Let's now go to the Input (Parameters) tab of the Strategy Tester.

In this field we can put the setup necessary for the realization of our
backtests. At the moment, what interests us is just the Value column,
highlighted in low red. The other columns (highlighted in green) are used as
parameters for EA optimization.

In the Agents tab we have information regarding the computational power of

the machine that we are using to do the backtests and optimizations. A
backtest is usually computationally inexpensive. However, when we are
doing optimizations (many backtests are required) we will need a lot of
resources and therefore this Agents tab will be important to us.
We can make a log to use multiple processing cores in the cloud and thus
increase our computing power. The use of these extra cores is very
affordable. With a few dollars we can make hundreds of thousands of
optimizations.

After setting the parameters correctly for the backtest, simply go back to the
Settings tab and hit the Start button.
7.2. Analyzing the Backtest

As soon as you press the Start button, the following window appears:

So we have a screen to visualize what is happening with the backtest. We can

control the speed with which new candles (candles) appear, from the
highlighted buttons (red) of the figure above.
You can also have the prices of each candle and indicator (green area
highlighted), just place the mouse over the desired element on the screen.
In the Toolbox we have four tabs (Trading, History, Operations and Journal)
where we can track records regarding our trades operations of our EA.
In the Toolbox in the Operations tab we can follow the records with each of
the sales and purchase operations. Something very useful is that we can give
two clicks on some of this operation and the graph will automatically show
the operation in question.
The Toolbox History tab shows a bit more detail. We can quickly see
profitable (green) and lossy (red) operations from colorful highlights.

Finally, the Operation tab shows the progress of each particular negotiation.
We have the information regarding the entry price, our available financial
margin and how the trade is progressing, whether with gain or loss.
7.3. Performance Charts

During the backtest or after the end we can follow the evolution of our capital
in the Graph tab that appears in the Strategy Tester.

However, after the end of the backtest, a new tab called Backtest appears.
This tab provides a summary of the operational statistics of our strategy. In it
we can see the profit Factor of the operation (ratio between gross profit and
gross loss), recovery factor, drawdown or retraction, total trading, number of
short positions and bought with gains and losses, sharp index among others.
We can visualize bar charts informing about the schedules, the days of the
week and the month where we have more entries in the traders.
You can track the frequency of profits and losses: by EA operating hours,
days of the week, and months of the year. We also have information
regarding the maximum time set and the minimum duration of a position.
However, as we can see, MetaTrader 5 gives us a very thorough review of
our backtest.

--==*__*==--
Congratulations on getting here!! Thank you for your confidence in reading
this e-book. Sorry for the problems with the translation. Feel free to submit
suggestions, criticisms and errors in the texts and subjects covered in the e-
book.
Now that you've been able to create your first EA and make the backtests, it
means you have a new journey of study towards more sophisticated
strategies. It is worth remembering that the EA taught in this book is purely
demonstrative and offers no means to use as a suggestion of real investments.
If you want to know other books, with similar didactics, it is worth checking
the following:
How to create indicators and scripts in MQL5 (step by step).
How to Properly Optimize Investment Robots in MetaTrader 5
If you are interested in any of these books please contact me by e-
mail:[email protected]

This material was written thinking of you dear reader. It is good to remember
that we are all motivated by criticism or praise. If this book was helpful to
you in any way, whether or not you liked the didactics, please leave a
comment on the Amazon store. This kind of attitude is of extreme
importance so that future editions are even better.

Thank you very much!

[1] Obs: The information contained in this e-book does not represent or offer any
suggestion and / or expectation of investment.
[2] https://en.wikipedia.org/wiki/Bill_Williams_(trader)

[3] More information send an email to: [email protected] .

[4] By the same author ([email protected])