Is backtesting a trading strategy worth the time and effort? Given that most professional fund managers backtest their methodology then the answer is a resounding yes.
Backtesting may have limitations and it may not provide all the answers but it does provide valable learning and insight.
Be aware that backtesting is not live trading, so the circumstances and emotions are going to be different when real money is not being made or lost. However, it can help validate a trading approach and give you confidence to ‘go live’ based on what you have learned.
If your strategy has proved its worth in theory, then you’ll be more assured to act in practice when the next trade signal shows up.
Technology plays an important role too - backtesting software allows you to learn quickly. For instance, you can analyse six months or a year’s worth of Forex price statistics in just minutes.
You set the agenda based on your distinct trading approach – the backtesting software then tests the reliability of this strategy. The backtesting software accurately manipulates the price data and applies your trading rules to it.
Your trading strategy rules need to be specific and consistent, so to understand when to take a trade when pinpointed on a chart. Without specific rules that you can follow every single time you trade, it will be impossible to backtest your strategy.
What you need to back test
You will need price data or a charting package as well as backtesting software. There are no shortage of providers offering backtesting software but shop around as they often have different features/strengths.
Sometimes data available to backtest can be fairly limited (for instance one to three months on a five-minute chart).
In essence, backtesting involves moving one candlestick (time-period) at a time until you see a trade setup you would take under your trading strategy. Future price movements should always be hidden so you don’t see the result of your trade until after you have agreed to take it.
There are several problems you might encounter when you backtest your trading system, so you need to be aware of them. We’ll pick out a few and suggest possible remedies.
Postdictive error means you have used data only available after the fact to test your system. It is a common error when backtesting.
For example, a feature of your system may be the closing price. If so, the trade cannot be initiated until the day is over, otherwise this is a postdictive error. The way to avoid the postdictive error is to ensure that when you backtest, only information that is available in the past at that point in time is used.
Degrees of freedom bias
Essentially the degrees of freedom bias is when you have too many variables in your trading system. Often when a trading system has too many indicators, you can use many of them to predict the movement of the market during a specific time period efficiently. However, there are flaws because in the future the system is less accurate.
The solution is not to over-complicate things – the most profitable trading system is often simple in nature.
Dramatic changes in the market
None of us who trade have crystal balls but from experience we know that at some point in the future markets will behave erratically. Your trading system should be designed to work during these periods.
For instance, if a Brexit vote or a Trump tweet sends currency markets wild, your strategy should have factored such possibilities in.
You need to be prepared for the unexpected. So, if your backtesting points to a maximum loss of £4,000, assume a maximum loss of £8,000. Will your trading systems still stand up under these conditions?
Agree on an appropriate level of risk for each trade. If you decide to risk 1% on each trade, you should assume that at some point an unexpected event will happen, and subsequently your trade will not lose 1%, but 5%.
With a maximum risk level, what if you have several trades open simultaneously. If you are prepared to risk 1% per trade but you have six trades open simultaneously, does this mean that you are prepared to 6% of your account? Or perhaps you have you decided on a maximum risk level of, for instance, 4%. You may need to set maximum risk levels for those periods when you have multiple trades open.
You will also need to be clear on the maximum drawdown (the amount of money your trading system loses over a defined length of time) you are willing to accept.
If you lose 25% of your account will you stop trading? Or what if it is 50% or even 65%? The most effective way to plan for drawdowns is to backtest thoroughly to establish what sort of drawdowns your trading system experiences historically and then use that as a base.
Whatever your level or frequency of trading, a contingency plan always makes good sense. If you can’t access your trading platform for some reason but you want to exit a position pronto, have you a phone contact to make that trade asap?
Backtesting your trading strategy will not guarantee success; but it will allow you to learn from experience and to trade with greater confidence in the future.