Summary
Backtesting is an important step in optimizing the way traders participate in financial market activities. It can help traders understand whether current trading ideas and strategies are reasonable and whether they can bring potential profits.
So, what does backtesting a simple investment strategy look like? What are the considerations when testing trading strategies? Are there similarities between backtesting and simulated trading? We will answer all these questions in this article.
Introduction
Backtesting is a tool that traders or investors can use when exploring new markets and strategies. Backtesting can provide valuable feedback based on historical data and discern whether an investment idea is sound.
Regardless of the asset class being traded, backtesting eliminates the need for traders to risk their hard-earned money. Using backtesting software in a simulation environment, you can build and optimize specific approaches to the market. See below for details.
What is backtesting?
In finance, backtesting tests the performance of a trading strategy against historical data to weigh its feasibility. In other words, it uses past data to observe how well a strategy performs. If the backtest results are favorable, the trader or investor can proceed to implement the strategy in practice.
But what does it mean to have a good outcome? Backtesting tools are used in order to analyze the risk and potential profitability of a particular strategy. Then optimize and improve investment strategies based on statistical feedback to maximize potential returns. Sound backtesting also ensures that the strategy is at least feasible in a real trading environment.
Of course, backtesting platforms or tools can also effectively evaluate whether a strategy will be unfeasible or risky at certain times. If backtesting indicates poor trading results, the trading idea should be abandoned or modified. However, it is also important to take market conditions into account when testing. Once market conditions change, even the same backtest will have very different results.
From a more professional perspective, backtesting trading strategies is absolutely essential, especially algorithmic trading strategies (i.e. automated trading).
How does backtesting work?
The implicit premise of backtesting is that what worked in the past may work in the future. However, this is actually difficult to determine. What is profitable in a certain market environment may fail in another environment.
Backtesting with misleading data sets can also lead to unsatisfactory results. Therefore, it is necessary to find samples of backtest time periods that reflect the current market environment. This is particularly difficult to achieve due to the unpredictable nature of the market.
Before backtesting a strategy, it's a good idea to decide exactly what information you want to gain. How can the strategy be feasible? Conversely, how can one overturn personal assumptions? If foreseen in advance, the outcome will be less likely to affect individual biases.
Backtesting should include transaction fees, withdrawal fees, and other fees that the strategy may incur. It's also important to note that, like obtaining high-quality market data, backtesting software is quite expensive.
On that note, to obtain historical data from the Binance Futures platform, please fill out this request form.
And remember, backtesting is just testing. Similar to technical analysis and charts, there is no guarantee that a test will work even if it produces good results based on historical data.
Backtesting example
Let’s look at a super simple long-term strategy for Bitcoin.
Let’s take a look at our trading system:
We buy Bitcoin on the first weekly close above the 20-week moving average,
And sell Bitcoin on the first close below the 20-week moving average.
This strategy only produces a few signals per year. Let's look at the time period starting in 2019.
Bitcoin weekly chart from 2019.
This strategy produced five signals within the measured time frame:
Bought for ~$4,000
Sold for about $8,000
Sold for about $8,500
Sold for about $8,000
Buy @ ~$9,000
Therefore, our backtest results indicate that the strategy should have been profitable at that time. Does this mean it will definitely work in the future? Does not. This just means that looking back at this particular data set, the strategy should have been profitable at the time. This result can only be used as a rough baseline.
Note that we only looked at data older than two years. If you want to turn it into an executable plan, you need to go back to an earlier time period and test it with more price action.
Having said that, this is not a bad start. As long as the initial idea holds water, through further refinement we can build an investment strategy from it. Perhaps more parameters and technical indicators can be added to make the signal more reliable. It all depends on one's philosophy, investment horizon and risk tolerance.
➟ Want to start your digital currency journey? Welcome to buy Bitcoin on Binance!
Backtesting versus simulated trading
Now that we have a general understanding of backtesting, we have examined very simple investment strategies, and we also know that past performance does not reflect future results.
So how can we optimize a systematic strategy for current market conditions? We can experiment in real markets, just without risking actual capital. This practice is called "forward performance testing" or "paper trading".
Simulated trading (paper trading) is the simulation of strategies in a real-time trading environment. It is called "simulated trading (paper trading)" because although the transaction is recorded, no real funds are used. In this way, you can not only optimize the strategy, but also understand the performance of the strategy.
Sounds great, so where to start? Binance Futures Testnet is the perfect place to test strategies today without risking any capital. You can create an account in just a few minutes and test your strategies in a simulated environment, just like trading in real time on the real market.
We need to be wary of "cherry-picking" here, which means selecting only a certain part of the data to confirm a certain bias. The significance of forward testing is to bring the strategy into a preset real environment for verification. If the system gives operation suggestions, you can refer to them and execute them. If you choose trades that "look good" based solely on personal preference, the system's testing of the strategy will be ineffective.
Manual vs. automated backtesting
Manual backtesting involves analyzing charts and historical data and manually executing trades based on the strategy. Automated backtesting is essentially the same thing, except that the process is automated by computer code, for example, using a programming language like Python or specialized backtesting software.
Many traders use Google or Excel spreadsheets to evaluate strategy performance. These documents work similar to strategy tester reports and include various information such as: trading platform, asset class, trading hours, number of profitable and losing trades, Sharpe ratio, maximum drawdown, net profit, etc.
Simply put, the Sharpe ratio is used to assess the potential return on investment (ROI) of the strategy relative to risk. The higher the Sharpe Ratio value, the more attractive an investment or trading strategy is.
The maximum drawdown refers to the moment when the trading strategy performs its worst compared to the previous peak, i.e. the largest percentage decline in the portfolio during the analysis period.
Summarize
Many systematic traders and investors rely heavily on backtesting strategies. This is an essential tool in any algorithmic trader's toolkit.
But at the same time, interpreting the results of backtesting is not easy. Backtesting methods can easily be tainted with personal bias. Backtesting alone may not be able to establish a viable trading strategy, but it can be very helpful in testing trading ideas and keeping a finger on the pulse of the market.