Systematic approach to trading is executing all your orders in line with predefined parameters and rules. In other words, trader can not change his rules and parameters without validating the change and he has follow to his rules in his trading activity
To employ such approach, trader has to minimize his discretion and answer all possible questions regarding his future activities and test his assumptions by using scientific methods. Only this way, a trader can stand solidly against all his adversaries and achieve constant success.
In this article, I will discuss key issues regarding methodology of constructing of a trading system. I hope it will help all others like me. Since this is an article, I will only provide an overview and I strongly recommend below books for further details and any unanswered questions.
*The Evaluation and Optimization of Trading Strategies - Robert Pardo
*Trading Systems: A New Approach to System Development and Portfolio Optimisation - Emilio Tomasini and Urban Jaekle
*Teknik Analiz Mi Dedin? Hadi Canım Sen De! – Ali Perşembe (Only available in Turkish)
PHASES TRADING SYSTEM CONSTRUCTION
GUIDELINES TO PHASES
Step 1 Formulate your idea into a strategy
It is known that having two aces Texas Hold'em Poker gives you the highest percentage of winning chance. Having an edge is like entering into poker game every time when you have two aces. By completing this guidelines, we are trying to understand if our trading idea really assists us. In other words are we trading when we have good cards?
Write down your trading idea which you think can provide an edge in the markets.
Like poker, trading is not that simple. You should know when to bluff when to exit and answer all the questions before hand. For this reason, you need to transform your idea into clear understandable rules. Define your of entry and exit signals precisely. If you are going to test your idea, then you should define the idea into logical mathematical formula. There should not be a place for discretion in your formula.
Step 2 Find a test platform and define your trading system idea in testable form
You can use any number of software to test your strategy, excel, jforex, metatrader are just few of the options. It does not matter which interface is selected, you need to translate your rules into the interface language.
Every software has its advantages, but I strongly recommend Jforex mainly due its ability to process raw tick data and provide most accurate historical testing. Data quality does matter!
Step 3 Observe your strategy
In this stage, we are checking if our strategy is working properly.
If it is fundamentally flawed or performance of the strategy is not adequate, this means your thesis may need to be restated and Step 1 should be repeated.
On the other hand if your code is not working properly your code needs to be fixed and and Step 2 should be repeated
If everything is okay, you are ready to proceed into the testing phase. This means that you believe that you strategy which provides an edge in the markets and your code is working properly
Step 1 Multi Period Multi Market Test
We will be checking our strategy in many markets as possible and taking notes to improve our performance. Some strategies work better on some markets and the same is loser in another market. We are looking for an overall winner and analyzing why is it wins in some markets and why not in others.
For example, my Poseidon System was a loser with NZDUSD and success story EURUSD. After carefully examining the chart, I found out that NZDUSD even when it is trending is likely to make a pullback whereas EURUSD is more directional when trending with has less pullbacks. Due to the pullbacks of NZDUSD my strategy was a total loser and increasing stop loss was not helping as well. However now I know my Poseidon’s weakness to think about my future plans, I will not be using NZDUSD with my system (this way cutting my future losses) and I have a trading idea for NZDUSD which I am going to test in the future.
Also another vital thing is length of the testing period. At least 100 transactions should be conducted in each of tests to have results that are more or less reliable. Results of testing including less than 100 transactions are more likely to be random.
Step 2 Multi Period Multi Market Optimization
Optimization is very complicated procedure which aims to get best results by trying different parameters with your strategy.
However curve fitting and over fitting terms are the dangers of optimization. There is a chance that you can select parameters that only worked before and yielded great results but will not in the future. Understanding this dangers and avoiding them is crucial.
I am not going to explain how to optimize a trading strategy since it is lengthy process and it will differ greatly depending on your parameters that you want optimize. Please read the recommended books for further information and take Walk Forward Analysis method seriously which I am going to explain in the next step.
Step 3 Multi Period Multi Market Walk Forward Analysis
Walk Forward Analysis (WFA) is testing the final version of trading system with out of sample data. This method reduces the need for live testing and enable trader to start live trading with confidence.
Out of sample data means sufficient amount of data intentionally not used in testing and optimization procedures.
In simple, if you have used 2007 to 2010 data for your testing and optimization, you are testing optimized idea with 2011-2013 data. 2011-2013 data is out of sample data and testing with this data will show you how strategy fares with unknown market conditions.
If your strategy fails in out of sample data, either there is curve fitting or your strategy lacks something fundamental which you haven’t observed before. In any case, you have to restart from the beginning.
If your strategy succeeds to have good performance in WFA, you are ready for the live trading
REAL TIME TRADING PHASE
Constantly compare real time performance with test results and improve your system. At this stage your losses are more important than your winnings. Every loss provides you an opportunity to see the weakness of your system and improve it.
Set conditions for when you should stop trading and start reevaluating your system. For example, Emilio Tomasini and Urban Jaekle think that trading systems should not be always online. They suggest that if your system’s equity line crosses the moving average of system’s equity downwards you should stop trading with that system. Furthermore, trading system should be returned to the real trading if your system’s equity line crosses the moving average of system’s equity upwards.
Lastly, Money and risk management is important as your strategy do not under estimate to use adequate money and risk management depending on your portfolio and your system.
There is always room for improvement, hope you enjoyed. Happy Trading