Backtesting: Using the Software and Interpreting Historical Price Data

[RichardCox 10]

One of the most often discussed practices amongst technical traders is the activity of backtesting. But there are still many traders that are relatively unfamiliar with the process, so here we will discuss some of the details to get a better idea of what the term actually means. To start, it should be understood that backtesting is the process is using the price performance from prior time periods to test a specific trading strategy.

 

Here, traders will simulate trades with specific parameters (using relevant price feeds) and the effectiveness of a trading strategy can be gauged using real data. But when trading theories are backtested, it is important to remember that the results you receive depend heavily on the market conditions and the price movements that were present during the testing period.

 

Essentially, when the theory in question is backtested, the assumption is that markets will repeat themselves and that what has happened in the past will happen again in the future. So, if this does not happen (and market conditions change) there can be potential risks once the strategy is implemented using real money.

 

For example, if we back test a strategy for a currency like the Euro and include the time periods when the financial crisis of 2008 occurred, your results might be very different than if the time periods had only included times of more regular market volatility. So, when traders debate the validity of backtesting strategies, the main question can be seen when we ask whether or not past performance is indicative of what will happen again in the future.

 

Backtesting When Developing a System

 

Many proponents of technical analysis in forex will argue that backtesting is a critical component when developing an effective trading system. Since virtual trades are reconstructed with historical data, rules that are defined by an overall strategy will generate statistics that can be used to understand when a trading method is likely to work and when it is likely to fail. The ability to do this allows traders to optimize their approaches, find flaws in the original theory, and to accumulate confidence in a system before any real risk exposure is undertaken.

 

In the following sections, we will look at the application types that are used to conduct backtests, the types of information that is gleaned from these tests, and how to put this information to use when trading.

 

Application Tools and Data Types Accumulated

 

Proponents of Backtesting methods will argue that this approach can can provide traders with plenty of important statistical feedback when looking at the efficacy of a system approach. When looking at these statistics, some of the basics include:

 

• Net Gains and Losses - Typically expressed as a percentage, showing net profit or loss.
  
• Time Period - Describing the time frame when the testing was enacted.
  
• Asset Types - Showing the currencies that were included in the test.
  
• Volatility - Showing maximum upside and downside percentages
  
• Average Gain and Loss - Enabling traders to get a relative sense of their successful and unsuccessful trades.
  
• Risk Exposure - Showing the percentage of monetary capital that was exposed to market fluctuations
  
• Win Loss Ratios - Giving traders a sense of the number of gaining trades versus losing trades
  
• Returns (Annualized) - Shows total returns as a percentage for the year.
  
• Returns (Risk-Adjusted) - Shows returns as a percentage adjusted for risk levels.
  

 

When looking at most of the commonly available backtesting software programs, two screens will be viewed as most important. The first screen lets traders customize the settings for that will be used for each backtesting analysis. For example, these customizations will allow you to set items like time frames and spread costs. The first attached picture is example of one of these customization screens.

 

The second of these important screens will show actual backtesting results that will be contained in a report. Each of the vital statistics presented in the bullet list above will be shown in these fields. The second attached picture is an example of one of these screen within the backtesting program.

 

For the most part, these software programs are relatively similar and will include most or all of these statistical elements. There are examples of higher-end programs that will allow traders to add some increased functionality, allowing you to enact automated position sizing, trade optimization or other advanced features.

 

Key Things to Remember When Performing Backtesting Tests

 

Clearly, running a backtest can give traders some complicated statistics that might seem difficult to interpret at first glance. While some of these stats will be more important for traders than others, here we will look at some key things to remember when running backtests on your own:

 

• Always remember to consider the broader market trends that are seen during the tested time period. If, for example, a certain strategy works well during the period in which the 2008 financial crisis occurred, the same strategy might not fare as well when markets are showing less bearish (or less volatile) price behavior. This is the main reason why backtesting over longer term time frames (getting a variety of different market conditions) tends to be viewed as more of a valid exercise.
   
  
• Consider the behavior of the asset in question. Different currencies will also behave very differently in different market environments. So, a backtest performed for a high yielding currency might produce different results when a safe haven currency is used. This can be another factor that might distort your results if a wide variety of assets are not tested.
   
  
• Consider Market Volatility. Volatility is another extremely important factor to consider when testing your trading system. This is also important for accounts using leverage, as there can be margin calls that will change your stats if volatility is extreme.
   
  
• Always watch the average gains and losses figures. These numbers should also be taken with the win to loss ratio, as this can prove to be extremely helpful when determining your appropriate position sizes and aid in money management.
  

 

Combining All of the Relevant Factors

 

When looking at all of these factors, traders can then look to the annualized return figures, giving the trader an idea of how the system performs in the market when compared to other comparable approaches. To be sure, most traders will center most of their attention on annualized returns but traders should also consider the changes in risk level, as this will be another quantifiable factor. To make an assessment of these elements, traders should watch the risk-adjusted return, as this will establish parameters for different types of risk factors.

 

One all of these areas have been assessed, traders will be able to compare the relative strength of a trading strategy, as these results must outperform other investment approaches (and at less risk) in order to be considered a worthy venture. For many technical traders, backtesting is a vital part of the trading process but application customization must be given a great deal of attention as well. Failure to tailor your tests to the conditions established by the broker you will be using can lead to testing flaws and distortions in the data results.

 

Even amongst the practitioners of these methods, most would agree that Backtesting is not always the best way for accurately gauging the potential success or failure of a trading system (in all cases). This is because there will be many instances where strategies that were highly successful when tested in the past will not fare as well in the future.

 

But even with these possibilities, Backtesting is widely considered to be one of the most important parts of the process when traders are developing a workable trading system. When the process is conducted properly and interpreted accurately, traders will be aided when looking to optimize their approaches and to improve on their prospective strategies. This is one of the easiest methods for finding flaws in the process before making a real application of the strategy in the forex markets.

[/\/\ 10]

@RichardCox... thanks for your article on back testing. I have been testing various strategies however keeping it simple. I am using maximum loss and maximum profit, along with total profit & loss to test out feasibility of the market. However, I am not much worried about the total profit/loss since its rarely going to show profits due to the nature of the market. However, max loss and profit is key to know how well the market may perform.

Your thoughts?