Home

StrategyQuant User's Guide

1. 26 5 2 2 Building blocks 29 5 2 3 Strategy options 30 5 2 4 Genetic options 32 5 2 5 Money management 34 5 2 6 Robustness tests 35 5 2 7 Ranking options 37 5 2 8 Parts to improve 39 5 2 9 Parameters 40 5 2 10 Success criteria 41 5 2 11 Global options 43 5 3 Resu
2. 48 6 2 Optimization 53 6 2 1 Simple optimization 53 6 2 2 Walk Forward Optimization 59 6 2 3 Walk Forward Matrix 64 6 3 Portfolios 68 6 4 Strategy Editor 69 7 How to 71 7 1 Import history data from MetaTrader 71 7 2 Import new custom indicator 74 7 3 Export strategy from StrategyQuant and test or trade it in MetaTrader 81 7 4 Translate program to anot
3. 11 2 2 Genetic Evolution 12 2 3 Supported building blocks 13 2 4 Custom Indicators 14 3 Quick start with the program 15 3 1 Main concepts 15 3 1 1 Layout 15 3 1 2 Databanks 16 3 1 3 Working with files 17 3 2 Flow of work 18 4 Program modes 20 4 1 Data
4. StrategyQuant User s Guide Version 3 0 2 Last updated 9 8 2013 StrategyQuant User s Guide 2 Support If you ll have trouble understanding anything you need help or you simply have some question to ask related to the system remember your purchase includes also a support We are here for you you can contact us at www StrategyQuant com contactus Articles and forum on our website provide additional source of knowledge www StrategyQuant com articles www StrategyQuant com forum Copyright All rights reserved StrategyQuant software bonus strategies and content of this Manual is copyrighted You can use them only with valid license No part of this publication may be reproduced stored in a retrieval system or transmitted in any form or by any means including electronic mechanical photocopy recording scanning or otherwise without the prior written permission of the author 2012 2013 Mark Fric SonarBytes Ltd StrategyQuant User s Guide 3 Risk Disclosure Risk Disclosure Statement Trading any financial market involves risk This Manual is neither a solicitation nor an offer to Buy Sell any financial product The contents of this Manual are for general informational purposes only Although every attempt has been made to ensure accuracy the author does not give any expressed or implied warranty as to its accuracy The author does not accept any liability for error or omissio
5. Otherwise it contains equity chart for every robustness test simulation and table of important confidence levels computed using Monte Carlo analysis of simulations For information about how to use Robustness tests and how to interpret the results check the Robustness tests and analysis section StrategyQuant User s Guide 46 5 3 7 Strategy settings This screen displays current settings of the program versus settings that were used when the strategy was last time backtested generated The differences are highlighted with red color To quickly load settings from the strategy you can use the button Apply Strategy Settings This will load settings from the strategy and applies it to the selected program section StrategyQuant User s Guide 47 5 3 8 Source code Source code tab is where you ll get the result of the program For every selected strategy it generates selected source code Source Code Type There are three types of source code that can be generated Pseudo Code human readable pseudo code of the strategy You can see the strategy logic and you can use it for manual trading MetaTrader4 Expert Advisor the EA code for MetaTrader You can save the EA code to MetaTrader experts directory and you ll be able to run your new EA in MetaTrader MT4 Special EA Trade on Bar Open special version of MetaTrader4 EA it checks for signals and places trades only at the open of a bar This is valid not only for opening the trade but als
6. 10 Success criteria Walk Forward Success Criteria are related to Walk Forward optimization mode they are automatically evaluated by the program and give us quick result whether the strategy passes or fails the test Walk Forward optimization consists of periods of optimization part followed by run part 1 Overall Profitability overall profitability checks if the net profit result from optimization test is bigger than given value 2 Walk Forward Robustness this criterion measures robustness of walk forward runs For example if the optimization parts earned 10 000 and run parts earned 5 000 then the Walk Forward Efficiency is exactly 50 run parts earned 50 of what optimization parts 3 Consistency of Profits it checks how big percentage of run parts are profitable and compares it to given minimum percentage 4 Distribution of Profits this measures how big was the impact of every run to overall strategy profit For example if one period contributed to more than 50 of total strategy profit it means that the rest of periods performed quite poorly StrategyQuant User s Guide 42 5 Maximum Drawdown it checks if maximum percentage drawdown any run exceeded given maximum 6 Minimum trades it checks that there is at least minimum number of trades produced in every run If a run produces too little trades it is less statistically significant StrategyQuant User s Guide 43 5 2 11 Global options Don t store list of trades in
7. Manager 20 4 1 1 History Data 20 4 1 2 Custom indicators 21 4 2 Build strategies 22 4 3 Retest strategies 23 4 4 Improve strategies 23 4 5 Optimizer 23 5 Program Layout 24 5 1 Progress 25 5 2 Settings 26 5 2 1 Data
8. Options gt Strategy parameters If you check the first checkbox it will use the same parameters for long and short direction providing the rules are the same Click OK to store the settings and refresh the source code StrategyQuant User s Guide 55 Now you can see that the strategy now contains only parameters pEMA_1 pEMA_2 that are used for both directions Step 2 Setting optimization values To set up values that will be optimized we have to go to Settings gt Parameters Here you can see the list of all strategy parameters that are available for optimization Optimization simply means trying different values of input parameters For every parameter you want to optimize you have to check the line of the parameter and choose Start Stop and Step values The optimizer will iterate the value from Start to Stop taking Steps StrategyQuant User s Guide 56 Original value is also configurable it will be used to retest the original strategy You can use this value to compare performance of new results with the original settings The Number of tests value shows us how many tests have to be performed to test all the combinations of the values Note It is possible that your parameters table will contain much more parameters it could look like this This is another powerful feature of StrategyQuant It allows you to optimize not only strategy parameters but also other trading options such as how many trades to take per d
9. for us Select your two new records hold CTRL while clicking on rows to select multiple rows at the same time and then choose Data import gt Automatic Note Before running automatic import please close your MetaTrader otherwise the import will fail StrategyQuant User s Guide 79 Program will now automatically start MetaTrader and initiate custom indicator computation data export and its import to StrategyQuant All this is automatic you don t need to interfere with it in any way When the process finishes you ll see that the indicators now contain imported values Step 4 Using your new custom indicators Steps 1 3 have to be done only once for every new custom indicator and symbol timeframe output parameters you want to add to StrategyQuant When the indicator data are successfully imported you can use them in your building blocks StrategyQuant User s Guide 80 To use your new custom indicators in your strategies you have to check them in Building Blocks gt Custom indicators table Note The custom indicators are valid only for the symbol timeframe they were computed in our example only for EURUSD_fhdb H1 If you want to use them for another symbol or timeframe you have to repeats steps 1 3 StrategyQuant User s Guide 81 7 3 Export strategy from StrategyQuant and test or trade it in MetaTrader When you generate some strategies and find the ones you would potentially like to use in rea
10. indicator is an oscillator check this option Oscillator middle value for oscillators put their middle value the value around which they oscillate here Parameters parameters are a complete list of parameters together with their types and default values for the indicator If you define the custom indicator manually you can check its parameters in MetaTrader Output values these are the values indicator outputs If we ll check the screenshot of Keltner Channel indicator below you can see that it outputs three values in Data window and they correspond to upper middle and lower line on the chart StrategyQuant User s Guide 77 Click OK on the dialog and your new custom indicator is defined Step 2 Define custom indicator data Now that we defined new custom indicator definition we have to load the custom indicator data The definition that we made in the first step is only a description of the indicator To be able to use the indicator in your strategies you have to configure which parameter and output values it should use and load data for this indicator computed from MetaTrader Click on the Add new button in the Custom Indicators data table Here we have to choose our indicator Keltner Channel and choose which of its output values we want to use We decided for upper line If you want to use also middle or lower line you have to follow this step for each output you want to use Then you also have to choose the Symbol and Tim
11. is used during genetic evolution to compute the rank of the strategies so that the program is able to compare them Out of Sample this part of the data is used to verify if strategy really works as expected also on data that it was not evolved Note Setting the Out of Sample period makes sense for Genetic Evolution mode because only in this mode the fitness results computed in the In Sample data are used to rank the strategies to find the fittest strategies The other modes Random Generation Strategy Testing don t use evolution so you don t need to split the data to two periods For these two modes you can move the Out of Sample splitter to the far right leaving Out of Sample period empty StrategyQuant User s Guide 27 Test parameters Testing precision testing precision means how is the price data simulated during the backtest It is usually sufficient to use Selected Timeframe mode for Market orders or Tick simulation mode for Stop Limit orders You can also use faster mode for generation and then slower mode for strategies retest Trade On Bar Open In this mode the system will check for signals and places trades only at the open of a bar This is valid not only for opening the trade but also for closing the trade on stop loss or profit target If the trade reaches its stop loss or profit target it is NOT closed immediately at this level but only at the open of the next bar If you ll use this mode you have to genera
12. number of days check in the Progress panel These values 204 days of run and 475 days of optimization were computed from the available history data because we specified we want to have 6 runs and 30 Out of sample run data Below this information we can check the results for each optimization and run period You can see that only 3 out of 6 runs ended up in profit We can also check the equity chart StrategyQuant User s Guide 63 Interpreting the results How should we interpret these results First of all it is clear that this particular strategy will not benefit from reoptimization at least not from the reoptimization using the periods we specified 6 reoptimizations in 5 years which is roughly reoptimization every 200 days It is possible that this strategy simply not robust enough and it will fail in the future But there is also a possibility that the strategy could be profitable if different more frequent periods of reoptimizations are chosen How can we tell what is the best reoptimization period for strategy This is where Walk Forward Matrix comes to play StrategyQuant User s Guide 64 6 2 3 Walk Forward Matrix Walk Forward Matrix is a powerful feature that can help you with two things 1 Find the optimal period for strategy reoptimization it will help you identify the best optimization frequency 2 Test the strategy for robustness if the strategy passes Walk Forward Matrix test it means that with the help of
13. of the program StrategyQuant User s Guide 86 8 Appendix 8 1 Ranking criteria Net Profit Net Profit is a total profit loss the strategy produced Drawdown Max DD Drawdown is the measure of the decline from a historical peak in running cumulative profit of the strategy Max DD is the maximum percentage drawdown of the strategy Stagnation Stagnation stagnation is a maximum number of days during which strategy stagnates which means it doesn t make a new high on the equity Obviously you d want strategy with as little stagnation as possible of Wins percentage of winning trades Annual Return average annual percentage return of the strategy If it is 30 it means the strategy makes 30 every year in average Annual Return Max DD important ratio between annual percentage and maximum percentage drawdown You d want this number to be as high as possible Avg Daily Monthly Yearly profit average profit for given period in Return DD Ratio a very good measure of strategy quality the higher the number the bigger are profits in relation to maximum drawdown Expectancy performance metrics developed by Van Tharp it is computed as percentage wins average win percentage losses average loss StrategyQuant User s Guide 87 R Expectancy performance metrics developed by Van Tharp it gives you the average profit value related to average risk R that you can expect from a system over many
14. power so the faster computer you have the more strategies will be generated and tested The most important component that affects the speed of testing is processor CPU StrategyQuant is able to use all processor cores so the best results will be achieved by using multi core processor such as i5 or i7 1 3 Installation StrategyQuant comes with standard setup wizard you just download and run the installation EXE file and then follow the steps in the installation wizard Important Please DON T install StrategyQuant to standard C Program Files directory It might not work correctly because Windows security settings don t allow the program to write to its data files Instead install it to any normal drive or directory on the disk like C StrategyQuant or C Trading StrategyQuant StrategyQuant User s Guide 11 2 How does it work StrategyQuant is a program it doesn t have the brain or experience of a trader and it doesn t know how to create a profitable strategy What it does is that it randomly combines available building blocks indicators prices etc to create new trading rules The resulting strategy is then tested on a history data to see if it is profitable Random generation is the foundation of StrategyQuant Strategies generated this way can be further improved evolved using Genetic evolution 2 1 Random generation of trading strategies A trading strategy in the initial population is constructed usin
15. powerful tool but it is not a magic box that will start making you money with a click on a button It has to be used in the right way to get the results Generating new strategies in StrategyQuant is only about 50 of the work The rest of the work has to go into evaluating the generated strategies to filter out the ones that are curve fitted or not robust enough It is up to you to evaluate your new strategies properly and know their strengths weaknesses and limitations before you put them to live trading It can easily happen that from all the profitable strategies generated by StrategyQuant only 1 out of 10 passes the evaluation and we can consider using it for live trading But the number of strategies we can generate is almost endless so even 5 10 from infinity is a pretty big number There are few steps to evaluate the strategies quality and measure how good they will be in real live trading Please read the Evaluating Generated Strategies section for more information StrategyQuant User s Guide 10 1 2 System Requirements StrategyQuant is not an EA it is a normal program EXE file for Microsoft Windows and it will run on all standard computers with Internet connection Minimum system requirements 1 2 GHz processor 512 MB RAM 500 MB hard disk space Windows XP Vista Windows 7 or Windows 8 operating system Computing power Strategy generation and backtesting requires a lot of processing
16. strategy Find new trading strategies that are not only unique but also non obvious Reduce the time required to build a strategy from weeks and months to minutes Improve your existing strategies Is StrategyQuant Right for You If you trade using automatic trading system called also robots or Expert Advisors or you plan to develop your own trading strategies then StrategyQuant can save you money and hundreds of hours of your time Some traders prefer to purchase an existing trading robot there is a multitude of offers especially for Forex While this can be an effective way by purchasing someone s forex robot you are usually purchasing a black box you don t know how it works what are the exact rules and you are at mercy of its creator with potential adaptations to changed market conditions StrategyQuant allows you to create your own trading strategies exportable to a plain Expert Advisor source code so you have full control over your strategy What If You Are a Manual Trader You can still use for StrategyQuant to generate the trading ideas You ll be surprised to find many profitable strategies based on relatively simple rules that you wouldn t think of Every strategy that is created by StrategyQuant can be exported to readable pseudo code with full description of the trading rules and can be traded also manually StrategyQuant User s Guide 9 What to Expect Please keep in mind that StrategyQuant is a
17. top strategies 10 100 or 500 will be stored in Databank what comparison criterion to use which strategies to dismiss automatically for example those with negative P L StrategyQuant User s Guide 17 3 1 3 Working with files Strategies in Genetic Builder are saved in its own proprietary file format with str extension that can be opened only by StrategyQuant If you find potentially good strategies you should always save them so that you can work with them later Running strategies in MetaTrader MetaTrader cannot read the strategy str files If you want to test or run your new strategies in MetaTrader you have to export the strategy to MQL source code Please check the section How To Export strategy from StrategyQuant and test it in MetaTrader Please note that exported MQ4 files are not readable by StrategyQuant so make sure you always save your strategy also as a normal strategy file str StrategyQuant User s Guide 18 3 2 Flow of work The general flow of work when generating new strategies can be described as follows 1 Configure data for backtest You can use the history data that come with the program or optionally import your own data in MetaTrader format Then set up the time range and timeframe you want to use If you ll use genetic evolution mode you should divide the data to In Sample and Out of Sample periods You can also use additional data or Robustness tests to automatically test the strate
18. used in custom conditions are IS In Sample result of In Sample part of data OOS Out of Sample result of In Sample part of data RT Robustness Tests result of robustness tests P Portfolio results for portfolio if exists Strategy Selection Criteria here you can choose the criteria you want to use to compute the total Fitness of the strategy You can choose from the predefined most used criteria or you can build your own complex fitness function based on multiple criteria each with a different weight Note that if you choose to combine too many criteria they might fight against each other without achieving what you expected Profit Factor criterion to maximize Profit Factor of a strategy You can find explanation for every ranking criteria here StrategyQuant User s Guide 39 5 2 8 Parts to improve Strategy consists of three parts Entry rules Order type and Exit rules Each of these parts can be improved independently This way you can for example look for better exit rules or generate a new rule for long entry while keeping the short entry rule from original strategy You have to check the part you want to improve There is also a possibility to improve long short parts independently Addition types add adds new condition to the rule replace replace the whole rule with new condition add or replace randomly determines if to add or replace Improvement process Improvement means that StrategyQuant will make
19. BEING MADE THAT ANY ACCOUNT WILL OR IS LIKELY TO ACHIEVE PROFIT OR LOSSES SIMILAR TO THOSE SHOWN StrategyQuant User s Guide 4 Software License Agreement This legal document is an agreement between you the end user User and Sonarbytes Ltd Author AGREEMENT By installing StrategyQuant Software copying the Software and or clicking on the I Agree button during installation you agree to all of the terms of this software license agreement Agreement If you do not agree with all of the terms of this agreement click on the I Do Not Agree button and or do not install copy or otherwise use the software INTELLECTUAL PROPERTY The Software and any associated materials are protected by copyright law The package is a proprietary product of the Author The Author retains title to and ownership in the copyright of the software program and the associated materials You acknowledge that the Author owns all rights title and interest in and to the Software including without limitation all Intellectual Property Rights Intellectual Property Rights means any and all rights existing from time to time under patent law copyright law trade secret law trademark law unfair competition law and any and all other proprietary rights and any and all applications renewals extensions and restorations thereof now or hereafter in force and effect worldwide You agree not to modify adapt translate decompile reverse engineer disassembl
20. HTML reports possibility to customize HTML reports produced by the program You can customize which charts you want to see in the strategy performance report StrategyQuant User s Guide 44 5 3 Results 5 3 1 Databanks Databank is a storage place where the generated tested strategies are stored with their results Each mode Build Test Improve Optimize has its own independent databank table For memory reasons Databank cannot keep unlimited number of strategies instead it stores the selected number of top strategies for example top 100 or top 1000 strategies The configuration how many strategies to store in the Databank and how to sort them can be set in the Settings gt Ranking Options Databank functions Every strategy result in the Databank can be viewed in the Results screen by using double click Load this button is available only in Retest databank it loads Strategy projects str files so they can be retested Save as Strategy Portfolio HTML report Source code this way you can save your strategy you should always save good strategies to a Strategy Project str so that you can work with them later Edit Parameters Trading rules here you can manually edit the selected strategy either the values of its parameters or the trading rules Delete Clear all this will delete selected or all strategies from the list Create portfolio it will combine selected strategies to a portfolio a special report that combines
21. R PERSONAL INJURY TO THE EXTENT APPLICABLE LAW PROHIBITS SUCH LIMITATION FURTHERMORE SOME STATES DO NOT ALLOW THE EXCLUSION OR LIMITATION OF INCIDENTAL OR CONSEQUENTIAL DAMAGES SO THIS LIMITATION AND EXCLUSION MAY NOT APPLY TO YOU TERMINATION This license will terminate automatically if you fail to comply with the limitations described above On termination you must destroy all copies of the Software in electronic or other form including any copies on backup tapes or other media Upon termination of this License for any reason you shall have no right to refund of the whole or part of any License Fee paid StrategyQuant User s Guide 6 Table of Contents 1 Introduction 8 1 1 What is StrategyQuant 8 1 2 System Requirements 10 1 3 Installation 10 2 How does it work 11 2 1 Random generation of trading strategies
22. a copy the original strategy and randomly generate new conditions that will replace or add to the selected strategy part Then it ll test the new strategy variation to check if the improvement resulted in better performance Strategies that have better performance that original strategy are then stored to databank StrategyQuant User s Guide 40 5 2 9 Parameters this settings screen is available only for Optimizer When you load a strategy to optimizer then you have to specify which parameters should be optimized in the Parameters screen Every parameter that you want to optimize has to be checked and you have to define its Start Stop and Step values The optimization engine will try every value from start to stop making steps Original value is there for comparison with original strategy performance Number of tests The value in the Number of tests shows you how many tests in total have to be performed If you ll choose to optimize more than one parameter the optimization engine will test every combination of all parameters If you ll choose too many parameters to optimize it can result to thousands or even millions of tests Optimization method There are two optimization methods available Brute force brute force methods tries every combination This method Genetic optimization for every optimization with number of tests bigger than 500 or 1000 you should use Genetic Optimization method StrategyQuant User s Guide 41 5 2
23. agnates for given number of generation Fitness is the strategy score If it stagnates for 5 10 generations it is usually a sign that evolution reached its potential or went into a dead end so it might be a good idea to just restart it StrategyQuant User s Guide 34 5 2 5 Money management Money management or position sizing specifies how many lots or shares are traded on each trade StrategyQuant contains three different money management options that can be used in the program and also later n real trading in MetaTrader Fixed size strategy will trade with fixed number of lots This is recommended setting when you generate new strategy because it gives you clear overview of strategy real performance Fixed amount strategy will risk a fixed amount of money for every trade This is basic money management without compounding It can be used to test real performance of strategies where Stop Loss is based on volatility ATR or if you want to compare the performance of strategies with different Stop Loss Risk fixed of account advanced money management that is recommended for real trading The strategy will risk a given of equity on every trade This is simple but very effective money management that will allow the strategy to increase number of lots as your account grows It is generally recommended to risk maximum 2 5 of account equity per one trade StrategyQuant User s Guide 35 5 2 6 Robustness tests Robustness tests are a
24. anslate reverse engineer decompile decrypt reverse engineer disassemble except to the extent applicable laws specifically prohibit such restriction or create derivative works based on the Software copy the Software except for back up purposes rent lease transfer assign sub license or otherwise transfer rights to the Software or remove any proprietary notices or labels on the Software DISCLAIMER OF WARRANTY THE SOFTWARE IS PROVIDED ON AN AS IS BASIS WITHOUT WARRANTY OF ANY KIND INCLUDING WITHOUT LIMITATION THE WARRANTIES OF MERCHANTABILITY FITNESS FOR A PARTICULAR PURPOSE AND NON INFRINGEMENT THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE SOFTWARE IS BORNE BY YOU LIMITATION OF LIABILITY UNDER NO CIRCUMSTANCES AND UNDER NO LEGAL THEORY TORT CONTRACT OR OTHERWISE SHALL AUTHOR OR ITS SUPPLIERS OR RESELLERS BE LIABLE TO YOU OR ANY OTHER PERSON FOR ANY INDIRECT SPECIAL INCIDENTAL OR CONSEQUENTIAL OR PUNITIVE DAMAGES OF ANY CHARACTER INCLUDING WITHOUT LIMITATION DAMAGES FOR LOSS OF GOODWILL WORK STOPPAGE COMPUTER FAILURE OR MALFUNCTION OR ANY AND ALL OTHER COMMERCIAL DAMAGES OR LOSSES IN NO EVENT WILL AUTHOR BE LIABLE FOR ANY DAMAGES IN EXCESS OF AUTHOR S LIST PRICE FOR A LICENSE TO THE SOFTWARE EVEN IF AUTHOR SHALL HAVE BEEN INFORMED OF THE POSSIBILITY OF SUCH DAMAGES OR FOR ANY CLAIM BY StrategyQuant User s Guide 5 ANY OTHER PARTY THIS LIMITATION OF LIABILITY SHALL NOT APPLY TO LIABILITY FOR DEATH O
25. ay or what should be the time range for trading These settings are normally a part of Strategy Options but you can also optimize their values If you don t want to use them and see them in Parameters table go once again to Tools gt Options gt Strategy parameters and uncheck the checkbox for Add parameters for strategy options StrategyQuant User s Guide 57 The last thing we have to configure is the data that will be used for testing We can choose EURUSD on H4 timeframe for this example Step 3 Running the optimization Now we are ready to run the optimization We ll go back to Progress screen and click on Start button The optimization engine will test all the possible combinations of selected input parameters and stores the results for each combination to the databank on the bottom StrategyQuant User s Guide 58 We can sort the databank by Net Profit and we can see that the best input values in terms of maximum profit are pEMA_1 5 and pEMA_2 10 We can also see the results of original strategy gray background that used pEMA_1 10 and pEMA_2 20 Interpreting the results Now we ve got input parameters that were optimized for our given symbol and timeframe What we ve really done is we found out what worked best in the past We have to be very careful because the parameters might be ideal for the history data but there is no guarantee that what worked best on history data will work also in the future It is called curv
26. be a result of different order of trades Randomly Skip Trades it will randomly skip trades with given probability In real trading you can often miss a trade because of platform or Internet failure or simply because you paused trading for some time This test will give you an idea how the equity curve might look like if some trades are randomly skipped Randomize Starting Bar this will test the strategy behavior when the testing starts on a different starting bar It is obvious that a good strategy cannot be sensitive to which bar you start the test Randomize Strategy Parameters every strategy uses parameters such as period of an indicator or constant that is used in comparison This test checks the sensitivity of the strategy to a small change of parameter value Probability of change is a probability that any parameter changes its value Max parameter change is the maximum percentage to which the parameter changes its value For example if you set Max parameter change to 10 then a parameter with value 60 can be randomly changed to a range 54 66 10 of its original value of 60 Randomize History Data one very common case of curve fitting is when strategy is too dependent on history data This option checks the behavior of the strategy to a change in history data The Probability of change sets for every bar how probable it is that open high low or close price will be changed The Max price change is a percentage value of the change in rela
27. d parameters on the run period At the end of run period 1 the system again runs simple optimization on a part of data marked as optimization period 2 It finds the best set of parameter values and they are again used for trading in run period 2 This continues until the period 6 which is also the end of history data used in test Walk Forward optimization simulates how you d work with the strategy during real trading you d optimize it on some historical data and then trade it with the optimal values After some time you d want to reoptimize it and let it trade again Analysis of walk forward results can tell you if the strategy is a good candidate for periodic reoptimization Walk Forward optimization example Step 1 Configuring walk forward runs we ll follow with our Simple optimization example we have to load our strategy and choose the parameter values to be optimized only now we ll choose Walk Forward Optimization option in the Progress screen StrategyQuant User s Guide 60 For this option we have to specify also the walk forward settings Out of Sample this means how much of the whole period is left for run If we ll set it to 30 it means that in each period 70 of the data are used for optimization and 30 will be used for trading using the optimized values Walk Forward Runs this means how many optimization runs there will be which means how many times we ll reoptimize the strategy It is also possible to specify the opti
28. dditional tests that can be applied to your strategy The idea behind robustness testing is to verify how well the strategy will behave when there are small changes in inputs history data or other components of the strategy Robust strategy should have good performance even when the input parameters are slightly changed or when some orders are skipped StrategyQuant allows you to use Robustness tests in a form of Monte Carlo analysis This means that the program will run a specified number of strategy tests with given random changes As a result you ll receive a number of output results and output equities one for each test You can then check these results to see how the strategy behaves when the inputs are changed which helps you decide if you can trade such strategy Monte Carlo Simulation Settings Testing the strategy with randomized input or history data provides only one alternative that can be compared to the original result In order to get more complete picture the robustness test can be repeated until we receive bigger set of different results You can specify number of tests in the Number of simulations variable The more simulations will be run the better statistical significance of the results but it comes at a price Each test takes its time so the more tests you run the longer it will take to completely backtest the strategy The best value of tests is between 10 and 100 Confidence level for reporting is simply a confidence level t
29. e desired number of trades Stability a special value in range from 0 worst to 1 best that measures how stable is growth of the equity chart StrategyQuant User s Guide 88 You can see three sample strategies on the picture above All of them finished with the same profit but Strategy 1 was growing almost linearly very stable Strategy 2 was growing with occasional big drawdowns less stabile and Strategy 3 was moving up and down very little stability of growth Stability is a value that is quite good in representing a quality of the trading strategy and it can be used as the only one or the main criterion to compute the total strategy Fitness Symmetry criterion to maximize strategy symmetry Symmetry value is in and it is measuring how much is the Profit Loss for Long direction similar to Short direction For example if strategy makes 600 on Long trades and 400 on Short trades symmetry in this case is 66 400 is 66 from 600 If the strategy makes the same profit on both directions the symmetry will be 100 If one of the directions produces loss or 0 profit the symmetry will be 0 Win Loss Ratio criterion to maximize the ratio of winning trades vs losing ones Average Win Average Loss criterion to maximize average win or minimize average loss per trade Average Bars in Trade criterion to minimize the average number of bars the trade is open Average Bars Win Average Bars Loss criterion to minim
30. e fitting usually the more parameters the strategy has the bigger is the danger of curve fitting There are two approaches to curve fitting make sure the strategy is robust and not optimizing its values at all make sure the strategy benefits from periodic reoptimization So the question is will periodical reoptimization improve the results of my strategy If yes how often should I do it StrategyQuant can give you answers to these questions using another of its advanced functionality Walk Forward Optimization and Walk Forward Matrix StrategyQuant User s Guide 59 6 2 2 Walk Forward Optimization Walk Forward optimization is generally a set of simple optimizations followed by trading using the best parameters It is a technique in which you optimize the parameter values on a past segment of market data then verify the performance of the system by testing it forward in time on data following the optimization segment and the process can be repeated over subsequent time segments How Walk Forward optimization works in walk forward optimization the data are divided into a configurable number of periods 6 in this example Each period consists of optimization part and run part The program starts with optimization period 1 It will run the simple optimization on optimization period 1 to find the best parameter values These parameter values are then applied to run period 1 strategy is tested by trading with the optimize
31. e or otherwise attempt to derive source code from the Software REGISTRATION This program is neither freeware nor public domain Use requires valid license Contact us at http www geneticbuilder com contactus for bulk licenses or discounted prices GRANT Author hereby grants you a non exclusive non transferable license to use the Software upon payment of the License Fee until the expiry date of the license if any Author makes no guarantee of the frequency value applicability or content of future updates or modifications to the Software The Software will only be made available to you in electronic form for download The requirement to pay a license fee does not apply to evaluation copies for which Author does not charge a license fee Evaluation licenses expire 14 calendar days from the date of this agreement unless otherwise agreed to in writing by Author On the date of expiry of the license User agrees to either purchase the Software at the list price in force at that time or to destroy all copies of the Software in electronic or other form including any copies on backup tapes or other media User s use of the Software shall be limited to use on a single hardware chassis on a single central processing unit as applicable or use on such greater number of chassis or central processing units as User may have paid the required license fee You may not permit other individuals to use the Software except under the terms listed above tr
32. ed for retesting already existing strategies You can retest the strategy on another historical data or timeframe run Robustness tests on the strategy try different options for example trading only during certain hours per day 4 4 Improve strategies this section can be used to improve existing strategy created in the Build step Improving strategy means replacing some of its part entry rule order type exit rule or adding new condition to make strategy better Improver works in a way that it randomly generates the selected part of the strategy while keeping the other strategy parts intact This way you can look for better exit rule or for additional entry rule condition that will improve strategy performance 4 5 Optimizer optimizer can be used to optimize strategy parameters to optimal values as well as for Walk Forward optimization and Walk Forward Matrix analysis StrategyQuant User s Guide 24 5 Program Layout The program is divided into several independent sections that correspond with the main functionality 1 Home starting screen of the program it contains sample settings news and helpful links 2 Data Manager here you can manage your history data and custom indicators 3 Build strategies section to build new trading strategies 4 Retest strategies section to retest existing strategies with different settings or on different symbols timeframes 5 Improve strategies section to improve exist
33. eframe for which these indicator data will be valid Data will be computed in MetaTrader for a given symbol and timeframe so they will be not valid for any other symbol timeframe To use this custom indicator on another symbol simply repeat this Step 2 with another symbol If we want to use automatic import settings recommended you also have to fill the path to your MetaTrader this is MetaTrader that will be called to compute the data for this indicator We also have to specify the exact symbol name in that MetaTrader It is because in StrategyQuant you can use any name like EURUSD_fhdb but in order to run an export on MT4 we need the exact symbol name that is used in MT4 The last thing is set the indicator parameters StrategyQuant User s Guide 78 Again computed indicator data will be valid only for one combination of symbol timeframe output parameters If you want to use the custom indicator with another set of parameters you have to define it multiple times When all this is done click OK to save new Custom indicator data We can repeat the same step for the same indicator and symbol but this time we ll choose lower output value This way we ll have new custom indicator building blocks for upper and lower values of Keltner Channel indicator Step 3 Loading custom indicator data Now we have two new custom indicators defined but they still don t contain any data We can use automatic import and StrategyQuant will do all the work
34. empts to moderate selection pressure over time so that it is not too strong in early generations and not too weak once the population has stabilized and fitness differences are smaller The Greek letter Sigma is used in statistics to denote standard deviation and that s what it means here too The standard deviation of the population fitness is used to scale the fitness scores so that selection pressure is relatively constant over the lifetime of the evolutionary program Stochastic Sampling is a variation of roulette wheel selection It ensures that the observed selection frequencies of each individual are in line with the expected frequencies So if we have an individual that occupies 4 5 of the wheel and we select 100 individuals we would expect on average for that individual to be selected between four and five times The individual will be selected either four times or five times not three times not zero times and not 100 times Tournament Selection is among the most widely used selection strategies in evolutionary algorithms It works well for a wide range of problems StrategyQuant User s Guide 33 At its simplest tournament selection involves randomly picking two individuals from the population and staging a tournament to determine which one gets selected Initial Population Generation for initial population you can choose two options Generate random population Initial population will be generated randomly using the predefined setting
35. es We will run 10 simulations and there will be 10 probability that any trade is skipped Now when you run build or retest these robustness tests will be performed for every strategy as a part of the building of testing process StrategyQuant User s Guide 50 Robustness analysis results Robustness analysis shows us 10 different curves plus the equity curve for original strategy without any trade skipped You can see that depending on the trades skipped the results of the strategy can differ significantly If you ll look at the table on the left of the chart you ll see the Monte Carlo analysis of the results StrategyQuant User s Guide 51 What do these values mean The first row displays values of Net Profit Maximum Drawdown etc of original strategy for comparison The rest of the rows display values at different confidence levels These numbers are a result of Monte Carlo analysis applied on our 10 random simulations For example values at 90 confidence level mean that there is 10 chance that Net Profit Drawdown will be worse than the confidence level values Values at 95 confidence level mean that there is only a 5 chance that Net Profit Drawdown etc will be worse than these values In this test we randomly skipped trades Monte Carlo analysis shows us that by skipping 10 of random trades our Net Profit can decrease from 3800 to 2500 and Maximum Drawdown can increase from 6 32 to 7 27 and there is on
36. ewing a portfolio you can switch between portfolio chart and charts of substrategies using a combo box on the top right corner of the chart StrategyQuant User s Guide 69 6 4 Strategy Editor Note Genetic Builder strategies have a fixed format and editor matches this format with 4 tabs for Entry Long Entry Short Exit Long if we want to close the trade using some rule other than Stop Loss or Profit Target and Exit Short Each tab has IF and THEN part in the IF part you define the conditions and in the THEN part you define what should happen once the conditions are true To CREATE new strategy go to menu File gt Create strategy or click on the icon on the toolbar To EDIT existing strategy just select it in the databank and click Edit StrategyQuant User s Guide 70 StrategyQuant User s Guide 71 7 How to 7 1 Import history data from MetaTrader Step 1 Export data from Metatrader Open your Metatrader and go to Tools gt History Center There open the currency you want to export for example GBPUSD and double click on 1 Minute M1 so that it is refreshed on the right side of the screen Then just click on the Export button and choose your destination file Note StrategyQuant supports import of only 1 Minute data it will compute the higher timeframes automatically Now we have the data ready to be imported to StrategyQuant Step 2 Create a new symbol in StrategyQuant We can import da
37. g a combination of price patterns technical indicators order types and other parts to form the entry and exit rules StrategyQuant can use all standard technical indicators and oscillators like CCI RSI Stochastic etc time values like time of day day of week and price patterns These building blocks are then combined using logical and equality operators and or gt lt etc to form an entry or exit rule In addition it supports different entry and exit order types market order limit order fixed profit target exit after X bars etc With all the possible combinations of rules and orders StrategyQuant is capable of generating literally trillions of different possible trading strategies The building process itself is completely random builder randomly picks different building blocks from the available pool and combines them to create entry rule order type and exit rule There are some validity constraints that ensure that for example price is not compared to time value etc The result is a completely new random trading strategy Of course not every randomly created StrategyQuant User s Guide 12 strategy is profitable but StrategyQuant can produce and test thousands of new strategies per hour and there are many profitable ones in this amount 2 2 Genetic Evolution Genetic Evolution takes the process of finding suitable trading strategies even further In this mode StrategyQuant first creates a number of rand
38. gy robustness 2 Configure settings Go through all the settings and configure strategy type indicators and order types to be used for trading rules Optionally use time constraints to limit trading to a certain time range 3 Configure Ranking options Ranking options allow you to select Strategy Selection Criteria which is how the best strategies are determined You should also set up Custom Conditions to filter only strategies that pass certain criteria It makes sense to dismiss all the strategies that have too little profit or trades or too small Profit Factor Return DD ratio or System Quality Number 4 Run Build Start the building process Depending on your settings you can let it run several minutes several hours or even several days The more time it will run the more potential strategies it will test The best of them will be stored into the databank 5 Evaluate generated strategies Go through the generated strategies and evaluate them You can either evaluate them visually by checking their equity chart or by sorting them by their parameters in the Databank Choose the best ones to pass to the next step and save them to StrategyQuant project file sqn so that you can work with them later Evaluation can consist also of Retesting the strategies on additional symbols and or timeframes or with different spread and slippage and comparing the results You should also run Robustness Tests The goal of strategy evaluation is to fi
39. hat is marked with a dark background in the Robustness analysis table see below Usual confidence level to use is 95 Types of Robustness Tests Below is an explanation of different types of tests you can use If more than one option is checked the tests are combined Randomize Trades Order this is the simplest test it randomly shuffles order of the trades This doesn t change the resulting Net Profit but it is very useful in examining different variations of Drawdown that can be a result of different order of trades Randomly Skip Trades it will randomly skip trades with given probability In real trading you can often miss a trade because of platform or Internet failure or simply because you paused trading for some time This test will give you an idea how the equity curve might look like if some trades are randomly skipped StrategyQuant User s Guide 36 Randomize Starting Bar this will test the strategy behavior when the testing starts on a different starting bar It is obvious that a good strategy cannot be sensitive to which bar you start the test Randomize Strategy Parameters every strategy uses parameters such as period of an indicator or constant that is used in comparison This test checks the sensitivity of the strategy to a small change of parameter value Probability of change is a probability that any parameter changes its value Max parameter change is the maximum percentage to which the parameter changes its value For e
40. her language 85 8 Appendix 86 8 1 Ranking criteria 86 8 2 Entry and exit types 90 9 Final Words 92 StrategyQuant User s Guide 8 1 Introduction 1 1 What is StrategyQuant StrategyQuant is a program that automatically generates new unique trading strategies for forex stocks or ETFs Using StrategyQuant you can find profitable trading strategies for virtually any market any timeframe and any chart type No programming or trading knowledge is required The resulting strategies can be saved as a MetaTrader 4 Expert Advisor with complete source code With StrategyQuant you can Build an unlimited number of unique trading strategies Develop strategies for virtually any market or timeframe Save your strategies as a MetaTrader Expert Advisor with full source code Eliminate the manual labor previously required when developing a trading
41. ing strategy by generating additional rule conditions 6 Optimizer with features of simple and Walk Forward optimization and Walk Forward Matrix 7 Settings area each section has its own independent settings area where you can configure the build test process 8 Results area here you can see performance of the selected strategy and get its EA code StrategyQuant User s Guide 25 9 buttons for starting pausing and stopping the process 10 databanks lists where generated strategies are stored storage area for strategies it contains all the built tested strategies 5 1 Progress Progress screen is the place where you can start pause or stop the building testing process it also shows you the log and memory consumption Before starting the build or test you should have the data and settings configured The results generated top strategies will be continually stored in the databank on the bottom StrategyQuant User s Guide 26 5 2 Settings 5 2 1 Data in Data you have to select the symbol timeframe and time range on which the strategies will be backtested Main data for testing StrategyQuant tests all the generated strategies on a given history data and this screen allows you to choose the symbol and timeframe to be used for backtests There is an option to choose also the testing period Date From Date To Out of Sample Period allows you to split the history data to two parts In Sample it
42. ions These settings allows you to specify whether Stop Loss and Profit Target should be mandatory in the strategy and what is the minimum and maximum of the SL PT values in pips You can also define the desired Risk Reward ratio Having defined SL PT in the strategy is the simplest and many times the most effective approach If you unselect the mandatory SL PT then the randomly generated strategy can but doesn t have to have fixed SL PT It is advisable to use different exit rule for example exit after X bars or exit rule if you uncheck this setting otherwise the strategy will have no way to exit the trade Lookback periods Here you can set up the periods and coefficients used by the GB when generating the trading rules Maximum Period for Indicators sets the maximum period that will be used for indicators Usually we don t use periods bigger than 100 300 in the trading system depending on the type of trading strategy so it is advisable to set the desired maximum period value Maximum Period for Price Patterns this is the maximum lookback period for price patterns Open High Low Close prices It doesn t really make sense to set it to more than 10 20 because the older price usually has no relation with the current market movement Minimum and Maximum ATR Multiple these are coefficients that are used in adaptive Stop Loss or Profit Target if it is based on Average True Range value For example setting Minimum to 0 5 and Maximum to 5 te
43. ize the average number of bars for the winning losing trade Degrees of Freedom in a trading system relates to the number of criteria that are used to filter price action and or volume and determine entry points The more criteria and variables used to determine entry timing points the fewer degrees of freedom the system will have and vice versa StrategyQuant User s Guide 89 Degrees of freedom are compute from strategy complexity and number of trades The simpler the strategy is the more degrees of freedom it will have For this property the bigger value is better Complexity measures the complexity of the strategy It is simply a count of all the indicators prices operators and other building blocks that are used in the strategy The higher this number is the more complicated the strategy is For this property the smaller value is better Both properties can be used also in the ranking to influence the Fitness function StrategyQuant User s Guide 90 8 2 Entry and exit types Profit Target PT is the opposite of the Stop Loss it can be used to take the profit once the price moves in your favor As in the example above you can set up a profit target to close the position once it makes for example 100 pips Exit After X Bars this is very simple exit condition If used the strategy will close the trade after given number of bars time periods on a given timeframe This is many times a simple but effective way t
44. l trading it is time to test them in MetaTrader StrategyQuant normally saves strategies in its own proprietary str file format which is nor readable by MetaTrader In order to test strategies in MT4 you have to export its source code in MQL format This is simple go to the databank and find the strategy you want to use Double click on it which opens it in the Result details window above the Databank There go to Source code tab and switch the source code to MetaTrader4 Expert Advisor This will load the MT4 code of the strategy StrategyQuant User s Guide 82 Click on Save to file button and in the file dialog find the folder where your MetaTrader is installed for example C Program Files Alpari MT In this folder you have to go inside the experts directory and save the strategy source code there So the full path of the file will be for example C Program Files Alpari MT experts ERUUSD_H1_Strategy 1 1456 mq4 Now the strategy is copied to Metatrader You can open MetaTrader now In Metatrader go to menu Tools gt MetaQuotes Language Editor or press F4 This will open the language editor On the right side of the editor you ll have a list of strategies that are in the experts folder Double click on our strategy to open it in the editor window and then click on Compile on the top toolbar The strategy will be compiled and now it is ready for backtest or running live Note Compilation warnings are normal Please
45. languages All you have to do is to create a new language file for your language in StrategyQuant lang folder There is an English lng file that contains all language strings used in the program All you have to do is to create a copy of this file rename it for example to Deutsch lng for german and then open it in plain text editor like Notepad and add translation for every string The file format is simple if is language string in English then equals sign and after that should be language file in another language For example version Your translated string Homepage Your translated string Get Support Your translated string etc When you created new language file just restart the program and you should see your new language in menu View gt Language Notes for translation If you use any special characters the file has to be saved in UTF 8 format For strings that contain special marks like d s these marks will be replaced by the appropriate value by the program You should use them in the correct places also in your translated text The length of language string should be not much longer than the length of original string Use abbreviation if needed The reason is that there is only a limited place on the user interface for the texts If you ll use too long text it might not fit to the label When you create language file for a new language share it with others Send it to us and we will include it to the next update
46. lls SQ that it can use minimum 0 5 ATR as the SL or PT value and maximum 5 ATR as the SL or PT value StrategyQuant User s Guide 32 5 2 4 Genetic options settings that affect genetic evolution An important thing is that there is not one ideal configuration You can experiment with genetic evolution settings to obtain results that you want Genetic Options Population size is the number of different strategies in one population It should be at least 50 the ideal value will be between 100 1000 or perhaps even more Max of Generations is the count of how many generations the evolution should be performed Mutation Probability is the probability that a strategy will be mutated it should be around 5 15 Crossover Probability is the probability that the strategy is selected for crossover operation Every generation there is a crossover of strategies performed during which the two parent strategies mate and produce two children The recommended value if 95 Selection type Selection is an important part of an evolutionary algorithm Without selection directing the algorithm towards fitter solutions there would be no progress Selection must favor fitter candidates over weaker candidates but beyond that there are no fixed rules Furthermore there is no one strategy that is best for all problems Some strategies result in fast convergence others will tend to produce a more thorough exploration of the search space Sigma Scaling att
47. lts 44 5 3 1 Databanks 44 5 3 2 Overview 45 5 3 3 List of trades 45 5 3 4 Equity chart 45 5 3 5 Trade analysis 45 5 3 6 Robustness analysis 45 5 3 7 Strategy settings 46 StrategyQuant User s Guide 7 5 3 8 Source code 47 6 Advanced functionality 48 6 1 Robustness tests and analysis
48. ly 5 probability that our values will be worse that This means that even if our strategy has Net Profit of 3867 Monte Carlo analysis shows us that by skipping just 10 of the trades the Net Profit can decrease and there is 5 chance that Net Profit will be lower than 2539 By looking at the higher confidence levels we can see that none of our tests had worse results than 2539 Because robustness tests are generated randomly equity charts and values in the table will slightly differ every time you retest the strategy Also the more simulations you ll run the bigger statistical significance of this test Robustness Test Example 2 Randomizing strategy parameters and history data In second example we ll run robustness tests that will randomize both input data and history price This time we ll run 30 simulations to get more statistical data StrategyQuant User s Guide 52 When the backtest finishes we can see that the chart shows a bigger variety of different equity curves The look at the confidence levels table shows us that at 95 confidence level the drawdown doubles and Net Profit is less than half of the profit of original strategy however the strategy is still profitable so it seems to be robust enough StrategyQuant User s Guide 53 6 2 Optimization 6 2 1 Simple optimization The idea behind an optimization is simple First you have to have a trading system this may be a simple moving average crossover for exam
49. memory If you check this option all the strategies that are stored in databank will have their trades stored in temporary files This greatly saves memory because history of list of trades consumes the biggest part of the memory for every strategy If trades are stored in temporary files they are loaded only when they are needed for example when you ll switch to list of trades or to equity chart Don t store expired and replaced pending orders if you check this options expired and replaced pending orders are not saved to the strategy history This is default setting you should uncheck it only if you need to investigate something with the expired orders Strategy parameters this tab allows you to set the global parameters behavior Strategy contains two set of rules one for long and one for short side Usually they are symmetric Use the same parameters for long and short rules Parameters used in these rules for example periods for indicator can be named by common names same for long and short rule or separate for every direction Unless you want to optimize the strategy especially for long and short rule you should leave it checked Add parameters for strategy options this setting applies to Optimizer gt Parameters settings When you ll optimize the strategy you can choose to display only strategy parameters or also parameters for general trading options such as time range for trade etc Parameters without and with strategy options
50. mization In Sample and run Out of Sample periods by exact days you can do it by checking Define specific number of days Step 2 Configuring optimization criteria We can optionally set up also optimization success criteria they are automatically evaluated by the program and they ll give us an exact result if our strategy passed or failed the Walk Forward Optimization We ll set up our criteria as in the picture below StrategyQuant User s Guide 61 To find more information about what each criterion means check the Success Criteria page Now we can go back to progress panel and click on the Start button Step 3 Checking the results Walk Forward optimization takes longer than a simple one because there a 6 or more optimization steps instead of just one When it finishes we ll see that we have only two results in the databank Original strategy and Walk Forward result We can see that Walk Forward result ended up in loss 976 When we ll double click on the result in the databank the strategy results will open StrategyQuant User s Guide 62 We can see that the strategy failed in walk forward test based on our success criteria On the top you have the overall result and in the Test criteria table below you can see the results broken down by each success criteria used Below this table we ll see the reoptimization period that was used We didn t specify the days exactly we could by checking the Define specific
51. n All examples are provided for illustrative purposes only and should not be construed as investment advice No representation is being made that any account or trader will or is likely to achieve profits or loses similar to those discussed in this Manual Past performance cannot be relied upon as being indicative of future performance The information provided in this Manual is not intended for distribution to or use by any person or entity in any jurisdiction or country where such distribution or use would be contrary to law or regulation or which would subject the author to any registration requirement within such jurisdiction or country Hypothetical performance results have many inherent limitations some of which are mentioned below No representation is being made that any account will or is likely to achieve profits or losses similar to those shown In fact there are frequently sharp differences between hypothetical performance results and actual results subsequently achieved by any particular trading system One of the limitations of hypothetical performance results is that they are generally prepared with the benefit of hindsight In addition hypothetical trading does not involve financial risk and no hypothetical trading record can completely account for the impact of financial risk in actual trading For example the ability to withstand losses or to adhere to a particular trading program in spite of the trading losses are material poi
52. n Close Trade Note on Stop Loss Profit Target Move Stop Loss to BE Both Stop Loss and Profit Target can have either fixed value in pips like 30 50 or 100 pips or they can be based on ATR Average True Range for example 2 ATR This means that the actual value of the SL PT will be 2 times the value of average true range at the time the trade was opened StrategyQuant User s Guide 91 ATR based SL PT have the advantage that they adapt to changed market conditions and changes of volatility Some strategies but not all of them can be more profitable when they use ATR based SL PT instead of fixed one To enable disable ATR for SL PT you can enable disable the Average True Range indicator in the Entry Building Blocks StrategyQuant User s Guide 92 9 Final Words Congratulations on completing the guide to StrategyQuant I hope you ll like the program and you ll use it as an important part of your trading StrategyQuant is very new program and it is underactive development This means you can look forward for new exciting features that will move the program functionality even further I m open to a new ideas or suggestions so if you are missing something in StrategyQuant or if you think something can work in a different way don t hesitate to let me know I wish you many successful trades Mark Fric P S I would love to hear your success stories so please let me know at mailto support strategyquant com
53. nd strategies that are robust that means they work in different conditions and don t break up when there is a small change in parameters or in price data or they miss few trades StrategyQuant User s Guide 19 Optional additional steps are 6 Improve the strategy You can try to improve the strategy in Improver You can try to apply different combinations of exit rules or additional conditions to entry rules in search for better performance After the improvement you should again run the new strategy variation through robustness tests to make sure it didn t lose its robustness 6 Optimize the strategy You can run simple optimization to find better combination of input parameters of your strategy You can also run Walk Forward optimization to find out if the strategy would benefit from periodical reoptimization As the last step you can run Walk Forward Matrix analysis to determine the best reoptimization period There are articles on our website that describe more detailed flow of work http www strategyquant com articles getting 20started 20with 20StrategyQuant http www strategyquant com articles strategy 20building 20process StrategyQuant User s Guide 20 4 Program modes The program is divided into four independent data manager sections 4 1 Data Manager a place where you can manage history data and custom indicators used in the program 4 1 1 History Data This screen is where you can manage crea
54. note that there are some compilation warnings on the bottom These warnings are normal and they don t influence the strategy work There are simply some functions that are not used in the strategy and MetaTrader is informing you about that StrategyQuant User s Guide 83 Now that the strategy is compiled it is ready to be backtested You can close the MetaEditor go to the main MetaTrader screen and open Strategy Tester This will open the Strategy Tester dialog on the bottom and you can run the backtest Make sure you select the correct Expert Advisor Symbol Timeframe and Date From and To and then click on the Start button The test will start and after a while you ll get the results StrategyQuant User s Guide 84 Explanation of small differences in backtests If you ll compare test results in StrategyQuant and in MetaTrader you ll see that on some cases the backtesting results are not the same The results can differ slightly or significantly depending on the type of strategy Backtesting algorithm used in StrategyQuant is very accurate but it is not exactly the same algorithm used in MetaTrader so it produces slightly different result The important thing here is to understand that both testing algorithms are only approximations one isn t superior to the other StrategyQuant User s Guide 85 7 4 Translate program to another language The program is by default in English but it supports translation to another
55. nts which can also adversely affect trading results There are numerous other factors related to the market in general or to the implementation of any specific trading program which cannot be fully accounted for in the preparation of hypothetical performance results All of which can adversely affect actual trading results U S Government Required Disclaimer Commodity Futures Trading Commission Futures Currency and Options trading has large potential rewards but also large potential risk You must be aware of the risks and be willing to accept them in order to invest in the futures and options markets Don t trade with money you can t afford to lose This is neither a solicitation nor an offer to Buy Sell futures or options No representation is being made that any account will or is likely to achieve profits or losses similar to those discussed on this web site The past performance of any trading system or methodology is not necessarily indicative of future results CFTC RULE 4 41 HYPOTHETICAL OR SIMULATED PERFORMANCE RESULTS HAVE CERTAIN LIMITATIONS UNLIKE AN ACTUAL PERFORMANCE RECORD SIMULATED RESULTS DO NOT REPRESENT ACTUAL TRADING ALSO SINCE THE TRADES HAVE NOT BEEN EXECUTED THE RESULTS MAY HAVE UNDER OR OVER COMPENSATED FOR THE IMPACT IF ANY OF CERTAIN MARKET FACTORS SUCH AS LACK OF LIQUIDITY SIMULATED TRADING PROGRAMS IN GENERAL ARE ALSO SUBJECT TO THE FACT THAT THEY ARE DESIGNED WITH THE BENEFIT OF HINDSIGHT NO REPRESENTATION IS
56. o close the trade when the market moves sideways or when we want to lock in the profit Move Stop Loss to Break Even another simple but quite powerful condition It tells the strategy to move protective Stop Loss to the entry price break even when the given profit is reached This will not close the trade so the strategy will be still in position but there is zero risk after this move Because SL is at the entry price we ll not lose anything even if the trade reverses and goes against us This condition is evaluated only on open bar not on every tick Profit Trailing a simple trailing stop that trails the Stop Loss specified distance from highest achieved profit This condition is evaluated on open bar not on every tick Stop Trailing more advanced trailing stop that can use a value of indicator or price Open High Low Close to trail the Stop Loss For example the rule can trail SL at Lowest 20 20 pips This condition is evaluated on open bar not on every tick Exit Rule Price Operators Indicators it is a special type of exit where strategy has not only rules for entering the trade Entry rules but also rules for exiting it Exit rules Exit rules look as same as entry rules it is a combination of all the selected building blocks indicators operators price values Example of entry and exit rule pair Entry Rule if CCI gt 0 Enter at Market Exit Rule if RSI crosses above 100 and we are in long position the
57. o for closing the trade on stop loss or profit target If the trade reaches its stop loss or profit target it is NOT closed immediately at this level but at the open of the next bar EA generated this way will be 100 compatible with the strategy tested in GB using the Trade on Bar Open test precision Put values to parameters by checking this checkbox you can generate code where all the number constants like indicator periods comparison constants etc will be parameterized a parameter will be added for every constant This will allow you to optimize your EA in MetaTrader or test it with different parameter values StrategyQuant User s Guide 48 6 Advanced functionality 6 1 Robustness tests and analysis Robustness tests are additional tests that can be applied to your strategy The idea behind robustness testing is to verify how well the strategy will perform when there are small changes in inputs history data or other components of the strategy Robust strategy should have good performance even when the input parameters are slightly changed or when some orders are skipped Types of Robustness Tests Below is an explanation of different types of tests you can use If more than one option is checked the tests are combined Randomize Trades Order this is the simplest test it randomly shuffles order of the trades This doesn t change the resulting Net Profit but it is very useful in examining different variations of Drawdown that can
58. om strategies which are used as the initial population in the evolution This initial generation of strategies is then evolved over successive generations using genetic programming technology This process imitates the evolution the algorithm chooses the fittest strategies using selected performance criteria in every generation and the group of fittest candidates is then used to produce new generation of trading strategies As in evolution this should result in better and better candidates in our case in strategies that are more profitable more stable or generally better in the selected performance criteria StrategyQuant User s Guide 13 2 3 Supported building blocks StrategyQuant currently supports the following components to build entry and exit rules Indicators Simple Moving Average Exponential Moving Average Weighted Moving Average Commodity Channel Index CCI Relative Strength Index RSI Stochastic MACD Bollinger Bands Qualitative Quantitative Estimation QQE Triple Exponential Moving Average Custom Indicators Average Directional Movement Index ADX Average True Range ATR Momentum Williams Range True Range Price Difference Highest Lowest Keltner Channel Parabolic SAR Ichimoku Price Values Open High Open Daily High Daily Heiken Ashi Open Heiken Ashi High Candle Patterns Doji Hammer Bullish Engulfing Low Close Low Daily Close Daily Heiken Ashi Low Heiken Ashi Close Shoo
59. omponents but narrow your choice to a smaller group of indicators or price values You can find description of all Entry and Exit nodes here StrategyQuant User s Guide 30 5 2 3 Strategy options These settings allow you to specify the properties of the generated strategies as well as testing conditions Market Sides You can choose to generate strategies that trade only to one direction Long or Short or to both directions which is standard You can also select that you want the entry or exit rules to be symmetrical If they are symmetrical then the rules for both directions are the same only reversed An example of symmetrical rule s Go Long if CCI gt 0 Go Short if CCI lt 0 As an alternative you can choose to use non symmetrical rules in this case the rules for Long and Short sides will be generated independently An example of non symmetrical rules Go Long if CCI gt 0 Go Short if RSI lt 0 and Momentum lt 100 This settings can be used for both entry and exit rules for example you can have symmetrical entry rules but non symmetrical exit rules so the strategy will effectively use for example different stop loss and profit target for Long and Short orders Trading Logic Trading logic defines the behavior of the strategy and constraints of the backtester Exit at end of day end of range if selected the strategy will close all position at the end of the day or end of the trading range if defined Thi
60. or form will be automatically filled We can see that StrategyQuant recognized indicator Return type as Price it has two parameters and three output values Return type is a type of return value indicator generates it is used by StrategyQuant to properly match this custom indicator to other building blocks in the program so that it compares price with price and not for example price with CCI value It can be either StrategyQuant User s Guide 76 Number if it is indicator such as CCI RSI MACD etc Price if the indicator value is price like moving average or Bollinger Bands Price range if the indicator value is price range difference between two prices such as ATR or Bollinger Bands Range Boolean if the indicator returns true false values zero nonzero in our case You can have boolean indicator for example to recognize candle patterns or to implement your own simple trading rules How to decide the proper return type generally if the indicator draws its lines on the same chart as price then its return type is price If it draws its lines into a separate window below the main chart then it is Number with the exception of special indicators like ATR that compute some kind of price difference or range Is oscillator Some indicators like CCI or RSI are oscillators which means that its value oscillates around some number like CCI oscillates around 0 or RSI oscillates around 50 If the
61. ose different steps for the parameter values We chose pEMA_1 to go from 5 to 20 in steps of 5 and pEMA_2 to go from 10 to 60 in steps of 10 Perhaps the Walk Forward result would be better if both values are optimized to range from 5 to 50 with steps 5 Another setting that influences Walk Forward result is our choice of Selection Criteria In our examples we chose to select the best strategy by Return Drawdown ratio Who knows maybe the Walk Forward optimization would have better results if we d choose Strategy Quality Number score or R Expectancy You can repeat the optimization with another Selection Criteria to check it out StrategyQuant User s Guide 68 6 3 Portfolios There are two possibilities how to create a portfolio 1 Use Additional data to build test strategy on multiple symbols or timeframes this will create a portfolio that will contain results of the same strategy tested on different data 2 Manually combine strategies to a portfolio this will create a portfolio of different strategies combined together trading on the same or different symbols amp timeframes Portfolio results are a summary results of all the strategies it contains For portfolios created manually 2 there is a special Portfolio result created in databank that can be saved to a file for later use Portfolio chart Equity chart for portfolio combines equity of the portfolio as well as equity for every strategy in the portfolio When vi
62. ou can click on any of the cell on Multi Walk Forward Matrix table to update the results for the selected combination Let s choose combination 30 40 on the bottom right If you ll look at the Test criteria table you ll see that it failed only in 2 out of 6 criteria in Walk Forward Robustness and number of trades We can switch the Display combo box to show us net profit StrategyQuant User s Guide 66 This is not bad at all the best result was achieved for combination of 30 Out of Sample and 40 optimization runs which corresponds to reoptimization every 40 days with history of 91 days The strategy then made profit 6731 while the original strategy without reoptimizations made only 2760 So it seems that frequent reoptimization does benefit the strategy We can look at the equity chart for this combination it is not bad the chart is not ideal but it is a big improvement from our original strategy chart that looked like this We can see that reoptimized strategy remained profitable for most of the time even during the periods when the original strategy was losing Final conclusion We saw that this particular strategy benefited from frequent reoptimization It still failed our success criteria but perhaps with some additional filter its efficiency can be improved further StrategyQuant User s Guide 67 Possible improvements There are several possible improvements of our Walk Forward analysis for example we can cho
63. parameter reoptimization it is adaptable to a big range of market conditions Standard Walk Forward optimization tests the strategy results if it is periodically reoptimized let s say every 300 days But how do we know how often it is the best to reoptimize the strategy We can only guess unless we ll use Walk Forward Matrix that will test various combinations of reoptimization periods Walk Forward Matrix example Step 1 Configuring Walk Forward Matrix We ll continue with our example where we left it in the Walk Forward Optimization part This time we ll choose Walk Forward Matrix option In the matrix we have to specify which combinations we want to test for Walk Forward Optimization All the other settings combination of parameter values to test and success criteria are already set by previous examples so we can click on the Start button to start testing Computation of Walk Forward Matrix is very time consuming because it has to run full Walk Forward optimization for every combination in the matrix It is better to not set too many steps at least if you want to have the results relatively quickly StrategyQuant User s Guide 65 Step 2 Checking the results When the optimization is over we once again get a new Walk Forward Matrix result in our databank Double click on it to load it to the Results screen We can see that all the combinations of WF optimization failed according to the success criteria we defined before Y
64. ple In almost every system there are some parameters indicator periods comparative constants etc that decide how given system behave The optimization means to test the system with different parameter values to find the optimal values of these parameters giving highest profit or best Return DD ratio Optimization example Step 1 Loading a strategy for optimization First of all you have to switch to Optimizer window and load the strategy you want to optimize For this example we ll use simple EMA_Cross strategy that goes long when faster EMA crosses above slower EMA and go short when faster EMA crosses below slower EMA After you loaded the strategy it is added also as Original strategy to the Optimization results databank You can double click on the Original strategy and then go to Results gt Source code to see its rules StrategyQuant User s Guide 54 Make sure you check the Put values to parameters checkbox so that you see that variables pLongEMA_1 pLongEMA_2 pShortEMA_1 pShortEMA_1 are used to store indicator parameters In our optimization we ll try to find optimal values of these parameters There s still one small problem We can see that the strategy uses different parameters for long and for short direction We can use it like this if we want to find optimal values independently for long and short side but for our example we d like to use the same parameter for long and short side We can do it in program Tools gt
65. s and building blocks Decimation coefficient decimation means that the system generates more strategies than needed for initial population and then chooses only the best For decimation coefficient is 1 it means there is no decimation For decimation coefficient 2 system generates 2 x population size strategies and then chooses the best half for the initial population For decimation coefficient 3 system generates 3 x population size strategies and then chooses the best third for the initial population Use systems from databank as initial population existing strategies can be used as initial population These strategies have to be loaded to the Initial population databank Initial Population Conditions when generating initial population you can specify additional settings to filter bad strategies you don t want to use For example you can configure it to dismiss throw away strategies where Net Profit is below zero or where number of trades is below certain level Evolution Management here you can setup the repeated evolution Top strategies from all the evolution runs are stored to Top strategies databank so repeated evolution is a good way to let computer run for several hours or even days and keep it running continuous evolution Start again when finished if checked it will restart the evolution when it finished reached maximum number of generations Restart when fitness stagnates for X generations if checked it will restart if fitness st
66. s way you ll have no position open overnight Maximum Trades Per Day you can limit maximum trades the strategy takes per day Maximum Total Trades you can limit maximum total trades to a given number It is usually good to use this limit and set it to some big number 1000 5000 depending on how many trades you expect during a tested period Replace Pending Orders this is an important setting that has big influence on Stop and Limit orders If set to true the stop orders are replaced with updated version every time there is new signal and order is still pending In the older versions of StrategyQuant formerly named Genetic Builder this setting was implicitly set to false and it was hidden StrategyQuant User s Guide 31 It seems that setting it to true leads to better and more successful breakout strategies and it is generally recommended Limit Signals to Time Range Time Range From Time Range To this limits the hours the strategy is checking for entry signal to a given time range If used in combination with Exit At End Of Range then all open positions are closed at the end of the range If you don t check the Exit At End Of Range then the strategy will not open new trades outside the trading range but the already opened positions will be not closed Exit At End Of Range Exit At End Of Day Exit On Friday three settings that manage whether trades should be closed at the end of range day or week Stop Loss amp Profit Target Opt
67. ta to already existing symbol but that will overwrite its data so it is better to define a new symbol Start StrategyQuant go to Data manager screen History Data and click on Add symbol StrategyQuant User s Guide 72 Fill out the symbol and full name check the default spread and tick value Pip Tick point is the value of pip It means what number is 1 pip is usually 0 0001 For JPY based pairs 1 pip is 0 01 Pip Tick value in is a value how much is one pip worth in money it is usually 10 for all the currencies Pip Tick Step is a value by how much one pip can move Virtually all brokers now use 5 digit data so the value will be 0 00001 or 0 001 for JPY based pairs Click OK and the new symbol will be created The symbol doesn t have any data yet but we are going to import them in the next step Step 3 Import data to StrategyQuant Now select your new row with the symbol and click on Import data button on the top This will open the import dialog Import dialog is configurable it allows importing data from various file formats Since we use data from MetaTrader choose MetaTrader4 as a Predefined File Format StrategyQuant User s Guide 73 Choose the data file and click on Start Import button This will start the import process It could take few minutes depending on speed of your computer and data size When the import is finished it will display information window and asks you to close the dialog Now
68. tabank Options The best strategies found are continually stored into the Databank It is not possible to store every strategy remember that StrategyQuant can create thousands of new strategies every hour so we have to specify how many strategies should be stored in the databank how they should be sorted to find out the best ones and which strategies should be thrown away Maximum Strategies to Store in Databank simply the maximum number of strategies remembered in the program Dismiss strategies that have 100 duplicate orders if checked strategies that have exactly the same orders as some strategy in databank will be dismissed It is recommended to check this if there is a strategy with duplicate orders it will be likely the same strategy with redundant rules Custom Conditions You can define rules to throw away strategies with bad properties For example it usually makes sense to throw away strategies that result in loss or these that have too little number of trades Custom conditions allow you to specify your own custom rules that will be evaluated for every strategy If the strategy matches any of the rules it will be dismissed This is useful to quickly dismissing strategies that result in loss or with too little number of trades There are separate values for In Sample and Out of Sample performance and for Robustness Tests if used or Portfolio if you used additional data StrategyQuant User s Guide 38 The abbreviations
69. te import modify the history data that are used for backtesting StrategyQuant already comes with 4 years history data for the major pairs provided by ForexHistoryDatabase com but if you have your own history data in Metatrader you can easily import your data Check the section Import history data from MetaTrader for more description StrategyQuant User s Guide 21 4 1 2 Custom indicators In this screen you can manage and import custom indicators to be used in StrategyQuant The screen is divided into two tables Upper table 1 contains list of all custom indicators that are defined in the system Bottom table 2 contains indicator data data for an indicator with given parameters output values and symbol timeframe It is important to understand that StrategyQuant is not able to run indicator code so the only way how to work with custom indicators in SQ is by importing their data Data are dependent on indicator parameter values as well as on symbol and timeframe on which it was generated To find out more about how to work with custom indicators check the section How to Import new custom indicator StrategyQuant User s Guide 22 4 2 Build strategies the heart of the program Here you can generate new trading strategies using different configuration options and building blocks The resulting strategies should be saved to StrategyQuant file str so you can work with them later Available build modes Genetic E
70. te a special version of EA that will also trade only on bar open to achieve the same results Selected Timeframe only together with Trade On Bar Open it is the fastest testing mode It uses only the main timeframe to simulate the prices This results in a very fast backtesting with very good accuracy However for Stop or Limit orders the testing accuracy might not be sufficient and you should try more precise mode 1 Minute data slower testing mode it uses minute data if available to simulate price changes during the testing Tick simulation slow testing mode it uses tick simulation within minute data to obtain more precise results This is the most precise testing that can be achieved without real tick data This testing mode is accurate enough for any type of orders Real Tick this mode uses real tick data if available for exact price simulation It is much slower than the other modes and it should be used for final verification of new strategies Additional data for testing you can test your strategies automatically on multiple symbols timeframes by specifying additional data To select additional data choose the symbol in the combo box and set its timeframe time range spread etc StrategyQuant User s Guide 28 The strategy is tested on additional data only if the row is checked This way you can test your strategy simultaneously on multiple symbols or on the same symbol and different timeframes or even on the same symbol and
71. ther program such as MT4 for this specific symbol and timeframe also in SQ Remember every custom indicator is defined for a specific combination of symbol and timeframe on which it was computed To use the same indicator on another symbol or timeframe you have to recomputed its data in MT4 export it and create a new custom indicator for this symbol timeframe combination You can find more information about custom indicators in these topics Manage Custom indicators How to Import new custom indicator StrategyQuant User s Guide 15 3 Quick start with the program 3 1 Main concepts 3 1 1 Layout When you ll first start the program you ll see the main screen as in the picture below The program functionality is divided into tabs on the left side These tabs are Home starting screen of the program it contains the sample settings news and helpful links Data manager Build strategies Retest strategies Improve strategies Optimizer StrategyQuant User s Guide 16 3 1 2 Databanks Databank is the most important concept you should understand when using StrategyQuant Doesn t matter which mode you use the best resulting strategies are always stored in the Databank There you can sort the strategies by its properties load and save them and when you double click on a row it will open the strategy test details in the results window above You can configure how many
72. timeframe and only a different spread slippage or precision Viewing the results of additional data tests If you select any additional data you ll be able to check the results for each additional symbol timeframe as well as result for whole portfolio strategy tested on main data all additional data summarized together in the Results section If you ll use additional data for testing you ll have also the opportunity to filter the resulting strategies by the performance of the whole portfolio in Ranking Options Custom Conditions StrategyQuant User s Guide 29 5 2 2 Building blocks Building blocks are the core components that are combined together to create rules for every trading strategy Entry Rules Building Blocks Entry building blocks can be divided into four main parts price data Open High Low Close technical indicators RSI CCI Momentum etc operators that are used for comparison and to combine the rules lt gt and or etc time constants Hour Minute of day Day of week simple predefined rules CCI gt 0 Stochastic lt 50 etc You can choose check each component that you want to use in the strategy so you can select your favorite indicators or choose for example only price data operators if you want to generate strategies based only on price Good practice According to our experience you can sometimes get better results if you don t check all the available c
73. ting Star Dark Cloud Cover Piercing Line Bearish Engulfing Operators Greater Lower Crosses Above Crosses Below And Or Addition Subtraction Multiplication Equals Not Equals Time Values Hour Minute Day of Week StrategyQuant User s Guide 14 Order Types Enter at Market Enter Reverse at Market Enter at Stop Enter at Limit Exit Types Stop Loss Exit After X Bars Exit Rule Price Operators Indicators Profit Target Move Stop Loss to Break even Stop Trailing Profit Trailing We will be continually adding new technical indicators and other features to the Builder If you have your favorite indicator you d like to see in StrategyQuant just let us know 2 4 Custom Indicators Since version 2 0 the program offers almost infinite flexibility of building blocks with custom indicators In contrast with build in indicators StrategyQuant doesn t know how to compute custom indicators they are defined by their data This will enable you to use virtually any of your favorite indicators in StrategyQuant There are no real limits you can use multi timeframe or multi symbol indicators generally every indicator that works in MetaTrader can be The way it works is that you have to export the values of the indicator computed on a specific symbol and timeframe to a file and then import this file as a custom indicator to the Builder This way you ll create new custom indicator with values computed in ano
74. tion to ATR Average True Range So if for example close price is randomly chosen to be changed ATR value is 10 pips and Max price change is 20 then the price can change by 2 pips StrategyQuant User s Guide 49 Monte Carlo Simulation Settings Testing the strategy with randomized input or history data provides only one alternative that can be compared to the original result In order to get more complete picture the robustness test can be repeated until we receive bigger set of different results You can specify number of tests in the Number of simulations variable The more simulations will be run the better statistical significance of the results but it comes at a price Each test takes its time so the more tests you run the longer it will take to completely backtest the strategy The best value of tests is between 10 and 100 Confidence level for reporting is simply a confidence level that is marked with a dark background in the Robustness analysis table see below Usual confidence level to use is 95 What is Monte Carlo Simulation Analysis Monte Carlo simulation is a problem solving technique used to approximate the probability of certain outcomes by running multiple trial runs called simulations using random variables By running multiple simulations with random robustness tests we can perform the Monte Carlo analysis on the results Let s run a simple simulation to test how the strategy is affected by randomly missed trad
75. trades You can find more at http www vantharp com tharp concepts expectancy asp R Expectancy Score standard R Expectancy doesn t consider the length of testing period and number of trades produced There is a difference is you ll make let s say 2000 per year by making 10 trades or by making 100 trades R Expectancy Score adds a score for trades frequency It is computed as R Expectancy averageTradesPerYear SQN System Quality Number performance metrics developed by Van Tharp it is the measure of the quality of a trading system You can find more at http www vantharp com tharp concepts sqn asp Standard interpretation of SQN is Score 1 6 1 9 Below average but trade able Score 2 0 2 4 Average Score 2 5 2 9 Good Score 3 0 5 0 Excellent Score 5 1 6 9 Superb Score 7 0 Keep this up and you may have the Holy Grail SQN Score System Quality Number Score just like in case of R Expectancy SQN doesn t consider the length of testing period and number of trades produced In fact it is more favorable for systems that produce more trades without considering the length of the testing period It is computed as SQN averageTradesPerYear 100 Number of trades simply a number of trades of this strategy in the backtest You can use this criterion to approximate some preset value you would like to achieve for example total 100 trades The value of this criterion will be bigger for strategies closer to th
76. trading results of multiple strategies You can then check analyze the results of whole portfolio of strategies Views Each row in the table represents one strategy plus the results of the strategy backtest number of trades total profit loss Fitness score etc are visible in the table columns Results are divided into In Sample and Out of Sample parts StrategyQuant User s Guide 45 All the columns in the databank are customizable you can choose which columns you want to see using views Export to CSV you can also export the whole table content to a CSV file 5 3 2 Overview Overview screen displays various statistics computed from the backtest data 5 3 3 List of trades Contains complete list of trades generated by the backtest If you backtested your strategy on multiple symbols timeframes you can switch between the results or check the list of trades for whole portfolio 5 3 4 Equity chart Displays equity chart of the strategy Stagnation marker shows the maximum stagnation period this is the longest period that it took strategy to make new high on the equity In case of portfolio it displays charts for every symbol timeframe tested plus equity of complete portfolio 5 3 5 Trade analysis This screen displays yearly performance of the strategy plus configurable small charts with trades per hour day of week day of month etc 5 3 6 Robustness analysis This screen is empty if you didn t configure any Robustness Tests
77. volution StrategyQuant first generates initial population of random candidates using the Random Generation mode and then uses genetic evolution process to evolve the population and produce better and better candidates with each generation The process ends when predefined number of generations is reached or when there s no further improvement Pros in theory it should lead to strategies better than the initial random generation this means that the already good strategies in the first generation can be further improved search for profitable strategy in the trillions of possible combinations can be more effective with the power of evolution Cons evolution can be slower sometimes the evolution can lead to the dead end so the generation should be watched the group of generated strategies is limited by population size Random Generation In this mode StrategyQuant continually generates and tests new random strategies one after another until it is stopped The top candidates based on predefined criteria are stored into Databank so you can review them later Pros faster and simpler than genetic evolution it will run until it is stopped so if you let it run for a few days it can generate and evaluate millions of strategies Cons once the strategies are generated they are not further improved StrategyQuant User s Guide 23 4 3 Retest strategies this section is us
78. we have the new data successfully imported into StrategyQuant and we can use them for tests or new strategies generation StrategyQuant User s Guide 74 7 2 Import new custom indicator Custom indicators are indicators that run under MetaTrader4 platform and that are not implemented in StrategyQuant StrategyQuant is not able to run indicator code so the only way how to work with custom indicators in SQ is by importing their data Note Custom indicators are computed in MetaTrader not in SQ To get the correct results you have to use the same history data in both StrategyQuant and MT4 It will not work to compute your custom indicator in MetaTrader with your broker data and then use them in StrategyQuant on data from other source Before using custom indicators synchronize your history data so that both SQ and MT4 use the same history Step 1 Define new custom indicator in StrategyQuant Go to Data manager gt Custom indicators and click on Add new in the Custom Indicators definitions table This will open new indicator definition dialog You can define your indicator manually but if you have a source code of your indicator mq4 file you can use automatic recognition Click on Select and find your indicator file it must end with mq4 extension We ll try to import Keltner Channel indicator to make an example StrategyQuant User s Guide 75 The indicator file will be loaded and parsed and all the fields in Add indicat
79. xample if you set Max parameter change to 10 then a parameter with value 60 can be randomly changed to a range 54 66 10 of its original value of 60 Randomize History Data one very common case of curve fitting is when strategy is too dependent on history data This option checks the behavior of the strategy to a change in history data The Probability of change sets for every bar how probable it is that open high low or close price will be changed The Max price change is a percentage value of the change in relation to ATR Average True Range So if for example close price is randomly chosen to be changed ATR value is 10 pips and Max price change is 20 then the price can change by 2 pips For information about how to use Robustness tests and how to interpret the results check the Robustness tests and analysis section StrategyQuant User s Guide 37 5 2 7 Ranking options When the strategies are generated every new strategy is backtested on a history data and the results of the backtest are then used to compute the Fitness rank of the strategy Fitness is number from 0 to 1 and it should reflect the quality of the strategy according to the given criteria In this screen you can configure how this Fitness value is computed Strategy Selection Criteria criteria how many top strategies are stored in the databank which strategies are stored to databank and which are thrown away Custom conditions Da

StrategyQuant User's Guide

Contents

Download Pdf Manuals

Related Search

Related Contents