Backtesting of trading strategies: features and nuances

Have you ever watched a chart and saw a familiar model of technical analysis, but were not completely sure how you should approach trading? Thousands of traders experience this feeling of uncertainty every day.

On the other hand, there are traders who are more prepared and know what their next step should be. Many of these recent traders have spent countless hours researching and exploring pricing models through history testing. And this allows them to stick to their trading plan with a higher level of confidence.

So what is testing trading strategies? This is the process of using a strategy tester based on historical price data. You can perform manual testing by printing exchange rate charts or by viewing charts online. In addition, you can program trading algorithms that will perform testing according to the specified parameters.

Regardless of how you decide to test your strategies, the process itself will help you analyze emerging market situations and will undoubtedly provide you with a certain trading advantage.

Manual testing of trading strategies

Manual history testing can be quite tedious and difficult, but it is a sure and proven method. However, this method alone is not effective enough and makes a large probability of errors. For example, if you look at a chart, it can be difficult to determine if the price actually generated a lower low compared to the previous price level.

Manual testing of a trading strategy will allow you to evaluate the viability of your trading idea. You can look at historical data to see if your ideas will work.

It’s best if you have data for five or ten years, especially if you want to check the daily or weekly strategy. If you are trying to find an intraday strategy, it would be better to use data from a couple of years to test your ideas.

A thorough analysis can include a lot of data, and finding reliable data here can sometimes be difficult. For example, if you analyze tick charts, you will need to evaluate 1,440 ticks for each day, which exceeds 1 million ticks over a three-year period.

Manual history testing can be quite tedious and difficult, but it is a sure and proven method.

Automated Testing

There are several ways to test your trading ideas. You can use the strategy tester to test the data yourself, or you can use special software.

There are many free quote providers that let you download historical data for daily or weekly time frames. Most of these data points will show opening, closing, high and low prices. You can upload this data to a spreadsheet such as Excel, which can then be imported to your testing platform.

If you want to test the strategy using intraday data, such as hourly, minute or tick data, you probably need to purchase data from a provider. The advantages of purchasing data from a supplier are that, as a rule, their data is already filtered and cleaned.

Any data you upload must be verified for accuracy. You must make sure that the data is correct, especially if you rely on highs or lows to enter a trade. Bad data can lead to erroneous results.

You must really understand your strategy and determine how historical quotes affect the test results. For example, if you look at daily data, you do not know if there was a maximum of a day before or after a minimum of a day. This can create a problem if your take profit and stop loss are close to the entry level, since your criteria can generate a false signal, even if the price movement did not occur in the required sequence.

Trading Strategy Testing Software

Many trading platforms today provide programming opportunities for robots and advisers that use technical indicators to establish a predefined set of rules for entering and exiting a transaction. These algorithms are easy to test on historical data, which allows you to see if the strategy worked in the past.

The strategy tester in MetaTrader is an example of an automated testing tool that has an integrated backtesting system. You can use the MQL4 language to build your own trading system. 

Creating an automated trading system (Expert Advisor)

There are several ways you can add a systematic approach to your trading. You can program the trading system yourself, using your own ideas and strategies. Or you can ask someone else to program the expert advisor using the strategies you created.

If your expert advisor uses common tools, such as moving averages or other technical studies, the most effective approach will be to use a trading platform, such as MetaTrader or Ninjatrader, to backtest your ready-made strategies.

Learning how to program may take some time, but for your convenience, some standard strategies, such as moving averages crossing or overbought and oversold levels, are already pre-programmed on most trading platforms.

If you decide that system programming is beyond your technical capabilities or requires special skills, hire professional MQL programmers who specialize in developing expert advisors. For example, this can be done here: https://nordman-algorithms.com/metatrader-programming/

Testing for stories gives you many benefits. You can determine whether your strategy meets certain risk criteria and whether it can work in different market conditions. Most importantly, you have the opportunity to see the result of trading on history before you risk your real capital. This does not guarantee profitable trading results in the future, but can help reduce the likelihood of potential losses.

Buying a ready-made trading strategy

Dozens of commercial trading systems are available on the market. Many of them were tested by developers, and some will show impressive results. Regarding commercially available trading systems, you should always assume that the yield curve may be too good to be true.

In many cases, these impressive systems are overly optimized, so they look very profitable based on historical data, but as a rule, when trading in real time, you will get only losses.

Therefore, be sure to test the trading system on a demo account or on historical quotes before using a strategy using real capital. 

Problems and Pitfalls of Testing

As already mentioned, one of the problems associated with testing and, therefore, buying a trading strategy is that there are methods that can be used to make the strategy look good on history but fail completely in real time. By choosing a yield curve or overly optimizing, you can create a system that has been tested in practice and looks very good for a certain historical period.

The system developer can slightly change the criteria that are used to achieve profitability. For example, you can test a trend following strategy by optimizing a moving average intersection system for two years. Once a result that looks good is found, it checks to see if the strategy has been working for a longer period. In most cases, the results will not be very impressive. However, you can find out much later.

In addition, many novice traders suggest that the trading system should have a very high percentage of profitable trades. With this in mind, an unscrupulous programmer can create parameters that can be adjusted, for example, to get an incredible gain of more than 90%. This may seem attractive to an inexperienced trader, but in the vast majority of cases, these trading systems use martingale. Ultimately, your losses will be many times greater than any series of profitable trades that this system will generate. 

Testing on stories is a great opportunity to determine if a trading strategy has the potential to work in the future. Keep in mind that just because the past performance of the system is positive does not necessarily mean that your strategy will work in the future. But that should give you more confidence. And this is the best we can hope for as traders. After all, we trade not by certainty, but by probability.