Banx broker
Advertisement
Share
View previous topicGo downView next topic
avatar
Broj poruka : 71
Points : 876
Date of Entry : 2015-05-25
Godina : 23
View user profile

Free Create trading robots for MT4 with StrategyQuant

on Fri Dec 23, 2016 11:53 am
Message reputation : 100% (1 vote)
Hi,
We have a two advanced EA programs that have been intended only for our clients :

  • EA Wizard (Web platform)
  • StrategyQuant (You need to download this program)



EA Wizard is designed for beginners ,with the following characteristics:

  • PLATFORM / Web, Mobile
  • NO PROGRAMMING SKILLS REQUIRED
  • NUMBER OF CREATED ROBOTS IS NOT LIMITED
  • EXPORT OF STRATEGIES INTO MT4


How to start using this free extra tool :

  1. You must be our client
  2. Deposited a minimum of $ 500



StrategyQuant is designed for For experienced traders ,with the following characteristics:

  1. PLATFORM / Desktop
  2. NO PROGRAMMING SKILLS REQUIRED
  3. NUMBER OF CREATED ROBOTS IS NOT LIMITED
  4. EXPORT OF STRATEGIES INTO MT4
  5. ROBUSTNESS TESTING
  6. TRADING STRATEGIES OPTIMIZATION
  7. CREATION OF TRADING ROBOTS USING MACHINE LEARNING


To use this great and free EA software is necessary:

  1. To be our client
  2. And deposited a minimum of $ 1,000 to your trading account
  3. Send your request to info@roboforex.com


To be continued!


Last edited by I_Robot on Mon Feb 27, 2017 10:57 am; edited 2 times in total
avatar
Broj poruka : 71
Points : 876
Date of Entry : 2015-05-25
Godina : 23
View user profile

Re: Free Create trading robots for MT4 with StrategyQuant

on Mon Feb 27, 2017 10:55 am

How to trade forex using RoboForex StrategyQuant software

WORKFLOW

This topic contains a description of the workflow that I use for profit when trading via RoboForex StrategyQuant. This workflow consists of several important parts. I am briefly introducing them on this page and later I am going to tell you about them in more detail:

Portfolio of strategies: What we are looking for
It is important to realize what we want to achieve with our strategies. We will discuss it in the first part

Import of data
The first important step is to import data correctly and set them accordingly. This is the most important step, because the result of backtests and robustness tests depends on the quality of input data. If the quality of the data is poor, the result of the backtest will also be poor and not reliable. Spend time on this part and learn how to process the data correctly.

Generating an initial package of strategies
Once the data is set up, it is necessary to generate an initial package of strategies. Ideally we build 1000 to 2000 strategies and then we do all the robustness tests and select those that are most likely to be profitable in the future.

Robustness tests
This is the crucial part. The robustness tests let you determine which strategies are most likely to be profitable in real trading and also how vulnerable they are to possible changes, such as a change of volatility, slippage, etc.

Walk forward matrix
Testing that uses the walk forward matrix method determines how reliable the strategy is and how much can be expected from it in the future. This is the most reliable test available.

Understanding the output of the strategy
This section describes what individual parameters of the strategy determine and how to recognize, what the subject of trade in the strategy is.

Application of strategies on a real account
This is the last step we are all looking forward to. This will be discussed in the last section: how to apply the strategy on a real account and how to use it.


PORTFOLIO OF STRATEGIES: WHAT WE ARE LOOKING FOR

RoboForex StrategyQuant's biggest advantage is that it allows to find good strategies easily and in a relatively short period of time. The aim therefore should not be to try to find one top strategy that earns hundreds of percent a year, but to find a portfolio of good strategies that complement each other. This allows to gain outstanding equity in real trading and long-term stable profits.

Let's have a look at an example. Here we can see one strategy from the portfolio, which trades on the EURJPY pair and timeframe H1:

Its equity is certainly not bad. At the risk of the maximum drawdown of 20 per cent it would earn 20 per cent per year. It is not bad, but we want more.

Therefore we build a portfolio of six high-quality strategies that complement each other:

This portfolio, consisting of six strategies earns 60 percent per year with the same risk of 20 percent.

Lastly, let's have a look at the portfolio of 15 strategies:

This portfolio achieves profit of 100 percent per year at the risk of 20 percent.

These examples show clearly that our goal is not finding one perfect strategy. We need to look for strategies that have good results that are robust and that have a great chance to be profitable in the future. Portfolio of such strategies gives us a good result with a very interesting profit.

Additionally, we can look for strategies on higher timeframe and do not have to rely on scalping in order to achieve high profits. Our portfolio will not therefore be vulnerable to slippages, spreads, etc., which is important for the reliability of the backtests. Scalping, which has an excellent historical equity, may not work in real market conditions.


To be continued!


Last edited by I_Robot on Thu Mar 02, 2017 10:38 am; edited 1 time in total
avatar
Broj poruka : 71
Points : 876
Date of Entry : 2015-05-25
Godina : 23
View user profile

How to correctly set data in RoboForex strategyquant

on Thu Mar 02, 2017 10:38 am
The topic of this article therefore is how to obtain and use high quality data.
This article covers:
• How to recognize high quality data
• How to download high quality data
• Tick vs. M1
• Data settings in RoboForex StrategyQuant
• How to import data to RoboForex StrategyQuant
• Setting a table for individual currency pairs.

HOW TO RECOGNIZE HIGH QUALITY DATA?
This is not an easy question to answer. In fact, the only proper way to assess data quality is to build strategies using this data, then run the strategy for several months and finally compare backtest results with live trading results. It is also worth noting that we have to use strategies with real trades, scalping strategies may not be a good choice.

Once you have a sample of high quality data, you can use it for reference and compare data from different sources with this sample. But it can’t be done without special software, which makes this process quite complicated.

From my experience I can recommend using data obtained from software Tick Downloader for strategy development, because the quality is relatively high. I discourage you from using the default MT4 data, because it is obtained from Metaquotes and not from a real broker, so it contains numerous errors and its quality is poor.

HOW TO DOWNLOAD HIGH QUALITY DATA?

I recommend using data downloaded by Tick Data Downloader.


  • TICK VS. M1


There is a popular opinion that backtest can’t be correct without using tick data. I believed this myth for a long time myself and I sticked to tick data, but then I made multiple tests that confirmed that even on M1 timeframe tick data is not needed. In fact, strategy for M1 timeframe can be tested using the 1 min data without significant differences in results.

Some backtesting deviation can appear, but it is not significant because statistically one trade has better results and another worse. So partial results may slightly differ, but the final equity is, in most cases, identical. In my experience I recommend using 1 min data. The 1 min data is almost 50 times smaller than tick data and you get much faster strategy development process when you use it.


THE DATA SETTINGS IN ROBFOREX STRATEGYQUANT

Correct data settings is the major issue for new program users. Very often users send me strategies with excellent parameters and brilliant equity, almost straight line. But if you get results like this, you probably have incorrect data settings. The most frequent problem is setting pip/step and pip/tick size, mostly on JPY pairs that are different from other pairs.

Let’s look at the specific settings. Window used for data settings looks like this:


Point Value in $:
Here you can set point value in USD. The Point is not the same as we know it from forex (move from 1.5001 to 1.5002). At this screen point represents price move from 1.5000 to 2.5000. There is rule for forex that the value is 100 000 units of base currency in the second place. For example for EURUSD it is 100 000 USD, for EURGBP it is 100 000 GBP, for AUDCAD 100 000 CAD, for USDJPY 100 000 JPY, etc. When you set the value to 100 000 for EURGBP, backtest will be inaccurate. Equity shape will be correct, but profit, max drowdown, average trade and everything where absolute value in USD is used, will be incorrect.

Pip/Tick Step:
It is the minimal move market can do. This is where users make mistakes most frequently. For example, for USDJPY it is 0.001, for EURUSD the value should be 0.00001.

Pip/Tick Size:
Another setting where users frequently make mistakes. This value is always by one decimal place shorter than Pip/Tick Step, so for USDJPY it is 0.01, for EURUSD it is 0.0001.


Data type:
This setting is easy, for MT4 data choose the first option, for data from other platform choose the second one. The reason you have to set this correctly is that MT4 data record the date and time at the beginning of the candle, other platforms record it at the end of the candle.

Default Spread:
Spread = difference between bid and ask.


Cost per roundturn:
Round turn cost, for both the opening and closing position.

DATA IMPORT INTO ROBOFOREX STRATEGYQUANT
After you set the symbol as explained in previous steps, continue with data import. Window for data import looks like this:

If you import data that the program does not recognize and its structure is not saved, then the cells with red border have to be set correctly.

The process is simple:

  • Choose the data and select row
  • Click on the Import data button
  • Popup window will appear.
  • Click Choose and select the data

When data is imported, program recognizes its format and sets all its parameters automatically. All you need to do is just click Start import and everything is then completed. Import can take several hours for tick data and a few minutes for 1 min data.

If you import data from different sources, proper input values have to be set:
If some rows in source file contain e.g. descriptions, then specify the number of rows to be skipped from the beginning of your data file that is being imported. The default is “0” i.e. no rows will be skipped.
For every column set appropriate content header. You can select values open, high, low, close, volume or unused (program skips this column).
Set corresponding date and time format. Make sure that month and hour letters in the timestamp mask are in upper case. If the date format is “22.10.2010”, then the mask is “dd.MM.yyyy”. If date and time are both in one column e.g. “22.12.2010 15:36”, then the mask is “dd.MM.yyyy HH:mm”.

Specify the delimiter that separates individual values in the data file, usually comma or semicolon. Table with settings for individual currency pairs. Lastly, you have to set the table with settings for individual currency pairs. Unfortunately, this settings can't be imported, so you have to set it manually. In the first column there are currency tickers, which are set for the second currency from each pair. For AUDCAD and EURCAD rows with “…CAD” value are valid.

Point values change over time. In the table below I show my setting. As mentioned earlier, it does not impact the equity curve but profit, max drawdown, average trade and everything, where absolute value in USD is used, would be incorrect. You can easily calculate them yourself.

Example of calculating CAD in USD.
Current value of currency pair USDCAD is 1.3244. That means that for 1 USD I can buy 1.3244 CAD. Value of one CAD is 1/1.3244 = 0.755 USD. Actual value in the table has to be 0.755*100000 = 75500 USD.

Formula for JPY is 1/119.5 * 100000 = 837

Reversely, the value $91600 means that in time, when I set this value, USDCAD value was 1.0917. The formula is 1/(91600/100000) = 1/0.916 = 1,0817. Which corresponds to the price levels in the first 9 months of 2014.


To be continued!
View previous topicBack to topView next topic
Permissions in this forum:
You cannot reply to topics in this forum