Walk Forward Pro Release Notes

New functionality and features are continually released into the Walk Forward Pro software. Take advantage by reading our release notes.

  • 22 May 2018

    Build 2.5.2.293

    Genetic Optimization Edition

    Genetic Walk Forward Optimizations

    Walk Forward Pro now supports MT5 ‘genetic’ optimizations to allow a wider range of parameters to be optimized in a much shorter time. To use the genetic option, ensure that the checkbox shown on the settings screen below is checked. Don't worry if you are an MT4 user - genetic optimizations are coming for you very soon...

    Genetic Optimization Settings

    Additional Selection Basis

    An additional ‘Selection Basis’ has been added as shown in the screenshot below (MT4 and MT5):​

    New Optimization Selection Basis

    These are defined as follows:

    Optimization Selection Basis Descriptions

    Additional Performance Metric

    A new selection performance metric has been added - 'Expectancy' (MT4 and MT5)

    New Optimization Performance Metric
  • 02 April 2018

    Build 2.5.0.283

    Walk Forward Pro 2.5 released - MT5 Multi Core Processing

    Version 2.5 of Walk Forward Pro brings full support for the multi-processor capabilities of MT5. Walk Forward Pro can now take advantage of optimizations that use the cores of the local PC in addition to local area network agents and MQL5 cloud agents. Because MT4 does not support multi-processor testing, this release has less relevance to MT4 users. MT4 users should still upgrade however, to take advantage of some improved chart capabilities that are also released in this version.

    The main settings that determine which cores/agents are used for optimizations can be found on the Settings screen.

    Under the ‘Optimization Settings’ heading you will see a number of new checkboxes. In order to use the multi-core capability, the first checkbox should be checked. If you un-check it, then Walk Forward Pro will operate in the old way where MT5 continually starts and stops, running each iteration separately. The default setting is ‘checked’, and you will need to leave it like this in order to use the new functionality.

    Multi-Core Optimization Settings

    There are 3 additional checkboxes to select which type of cores/agents you would like to use:

    • Local Cores (on the local PC where you run WFP)
    • Local Network Farm Agents (on other PCs in your Local Area Network)
    • MQL5 Cloud Network Agents

    You can choose to use any combination of these in order to perform your test.

    Using Local Cores

    This option uses the cores on the PC where Walk Forward Pro and MT5 are installed. The use of multiple cores means that optimizations will complete much quicker. The more cores your local PC has, and the more powerful those cores are, the quicker the optimization.

    Note that when using Local cores in MT5, you will only see a significant improvement in performance if your PC has more than one processor core. Many VPS starter packages only have a single core, and when this is the case, little benefit will be gained in terms of performance by using this option. Even so you should still use it because it will mean that MT5 does not need to continually start and stop during each optimization phase, saving a small amount of time.

    Using Local Network Farm Agents

    This option uses the cores on all the PCs that have been pre-configured as part of your MT5 local network farm. Note that this option will only work if you have already configured your network farm within MT5. You should test that this is working as expected in MetaTrader BEFORE attempting to use this option in Walk Forward Pro. In order to configure your MT5 Local Network Farm, we recommend you follow MetaQuotes guidance here https://www.mql5.com/en/articles/341

    Using MQL5 Cloud Network Agents

    In order to use agents from across the MQL5 Cloud Network, you must configure this in MT5 first (before using it from Walk Forward Pro).

    Follow MetaQuotes guidance here https://www.mql5.com/en/articles/341 to set this capability up in MT5. To use this capability, you must ensure that the connected MT5 application:

    • Is already signed into the MQL5 Cloud Network.
    • Has sufficient credit in your MQL5 Account to allow the cloud agents to be used.

    You must ensure that the cloud agents are working correctly when you perform a “manual” optimization in MT5 (not using Walk Forward Pro) before attempting to use this setting in Walk Forward Pro.

    IMPORTANT NOTE: Use of the MQL5 Cloud Network is not free. The more you use the cloud network, the more you will need to pay MetaQuotes for this service. It is solely your responsibility to monitor your usage and cost. Trade Like A Machine Ltd will not be liable in any way for any costs you incur. To better understand cloud network costs, see the MetaQuotes documentation here: https://cloud.mql5.com/en/faq/payments

    Combining Agents and Cores

    Note that any combination of Local Cores, Local Network Farm, and the MQL5 Cloud Network can be used simultaneously to achieve faster optimizations.

    Using Walk Forward Pro

    You now use Walk Forward Pro in exactly the same way as you did previously. Following the pre-assessment, if you are using the multi-processor capability you should notice that the status bar appears with text similar to:

    Multi-Core Optimization Status Bar

    You can also confirm this is working by viewing the ‘Agents’ tab of the MT5 Strategy Tester as it runs in the background:

    Metatrader Multi-Core Screen

    Additional charting improvements in this release

    With the new multi-processor/distributed-agent capabilities of Walk Forward Pro/MT5, it is now possible to run optimizations with many more iterations than previously possible. This is good news, but did mean that certain aspects of the WFP application needed to be re-considered.

    The main difference appears on the ‘In-Sample Optimizations’ screen. Previously, two bar charts were shown, representing i) the chosen performance selection criteria, and ii) The CAGR (Compound Annual Growth Rate) of the optimization undertaken (this old screen from version 2.0 is shown below).

    The bar charts did not work well for the high numbers of iterations that are now possible. With the help from a number of our beta testers we came up with something that we believe is far better and provides the user with much more intelligence. It also copes fine with large numbers of iterations. The first replacement is a scatter chart that means the data is much more readable than the bar chart (see below). Secondly we replaced the CAGR bar chart with a heatmap of the iteration performance. The heatmap facilitates the visualization of any 2 optimization parameters on an X-Y heatmap, providing insight and intelligence regarding the performance of the values tested (See below).

    Walk Forward Pro Heatmap and Scatter Chart
  • 18 December 2017

    Build 2.0.0.216

    Walk Forward Pro 2.0 release - The 'Machine Learning' Edition

    The machine learning module within Walk Forward Pro is designed to help traders achieve better and more robust results from their trading systems. It does this by assisting with the choice of settings that are most likely to produce an optimal walk forward optimization (WFO).

    As an example, it specifically addresses the following questions:

    • How many stages should be used in the WFO?
    • What ‘optimization to walk forward ratio’ should be used?
    • How many variables can be simultaneously optimized without over-optimizing (sometimes termed over-fitting) the system?
    • How much price data is needed to achieve statistical significance results?

    The complexity for the trader arises because the answers to each of these questions are inter-related and a decision for one of the questions will affect the optimal settings needed for the others.

    Although walk forward optimization is widely considered to be one of the best (if not the best) techniques available to test and optimize trading systems, the effectiveness of the process is often reduced by choosing inappropriate setting for the process. This leaves the trader with a system that will not reach its full potential, or even worse, with an over-optimized system that has little or no chance of producing consistent profits in a real-money account.

    That is where the Machine Learning Module provides a solution.

    Machine Learning Module

    An analysis of the statistical analysis of in and out of sample phases is first performed:

    In-Sample and Out-of-Sample Statistical Significance

    The Machine Learning Algorithms then provide suggestions for the Walk Forward Analysis settings in order to achieve more robust results:

    Machine Learning Analysis and Optimization

    Direct comparisons between the traditional and machine learning tests are shown for easy comparison:

    Machine Learning Optimization Comparison

    For a full description of the machine learning module, see our Machine Learning Manual

  • 01 October 2017

    Build 1.1.2.156

    Walk Forward Pro 1.1 release

    This build adds new functionality in the following areas:

    A - Three new checks in the Pre-Assessment Module:

    • Estimated statistical significance of the in-sample (IS) optimizations - Lack of statistical significance in the in-sample optimizations leads to a lack of predictive power to obtain the best parameters and can therefore be a major issue when back testing. This check warns the user up-front, when it is anticipated this will be a problem and suggests actions that the user can take to rectify.
    • Estimated statistical significance of the out-of-sample (OOS) back tests - Lack of statistical significance in the out-of-sample back tests leads to traders making decisions about whether a system performs well enough to trade in a live, real money account, based on insufficient intelligence and so is a major issue. This pre-assessment check warns the user up-front, before the walk forward analysis runs if it estimates this is going to be a problem, saving the trader time. It also suggests what the trader can do to improve the OOS statistical significance.
    • The final check makes an estimate of how long the full walk forward analysis process will take, and warns the user if this is an excessive length of time, giving you the chance of changing settings before the walk forward analysis starts

    B - A bug has also been fixed where an unexpected error was shown when attemping to select an EA before Walk Forward Pro had been connected to a MetaTrader instance.

  • 19 September 2017

    Build 1.0.1.153

    Support for MetaTrader Portable Mode

    A - This capability originated from a customer request and means that Walk Forward Pro can now work with MetaTrader in 'portable mode', if users wish to run in this way. The setting to achieve this is now shown on the 'Platform Setup' screen as shown below. You can read more about portable mode on the MetaQuotes website here

    MetaTrader Portable Optimization Mode

    B - Enhancement to error checking functionality and event logging

  • 05 September 2017

    Build 1.0.1.149

    Minor Enhancement and Bug Fix

    This builds adds the following:

    A - Pre-live results are now also saved indefinitely alongside the standard WFA results, meaning when users re-open past WFA tests, they will also see the pre-live metrics that were ascertained at the time of the run.

    B - One bug that caused a performance issue has been resolved.

  • 26 August 2017

    Build 1.0.1.142

    Internationalization

    Added capability so that Walk Forward Pro works when users who have set their decimal point to use a comma ( , ) instead of a period ( . ) in the Windows operating system regional settings.

  • 10 July 2017

    Build 1.0.1.129

    ​Post-launch tidy up

    Mostly bug fixes and cosmetic changes following the initial release and based on user feedback.

    Thanks to our small number of initial customers who've help us get things right.

  • 23 June 2017

    Build 1.0.1.123

    ​Initial product launch

    Welcome to Walk Forward Pro 1.0! We think this is the start of something big.

Like what you've read today? Then please consider sharing

Subscribe to our newsletter

By subscribing you are giving your consent to send you emails in accordance with our privacy policy

About Us

We are passionate about algorithmic trading, and about helping other algorithmic traders reach their full potential.

We help traders to develop robust trading systems that deliver results in live accounts.

Contact

Built in Yorkshire, UK
Proudly serving Europe and the World