n.b. at the end of this discussion I list some of the tools and resources currently available for testing forex strategies.
Backtesting is often criticised by those who claim it is either unnecessary and/or that you cannot guarantee results in the future based on analysis of past events.
It is important to acknowledge that backtesting is limited. You cannot guarantee future performance based on past results. However, lacking anything else to go on, it is a useful tool for putting a strategy through its paces before you risk money on it.
Backtesting means taking the rules for a trading strategy and subjecting them to historical data to see if it would have traded profitably in the past….
For example, let’s say we have a simple system based on the following (Obviously this is an extremely simple system which you would never trade. It is for illustration purposes only):
- If the daily candle closes lower than it opened, sell at the following daily open
- If the daily candle closes higher than it opened, buy at the following daily open
- Trail the stop behind the last daily candle each day
- Allow the market to take us out of the trade when our stop is hit
So now we take these rules of engagement and apply them to a given period in the past for which we have reliable data. We scroll through the data looking for trade entry signals for our system.
Having found one, we see whether the rules for trailing the stop behind the last daily candle would have resulted in a profit or loss for that trade.
And of course, we repeat this process for a given period of time until we are satisfied with the results, one way or the other.
This approach above implies manual backtesting. You can also automate the process, via an expert advisor or other such program that runs the system through its paces.
Some of the advantages of backtesting systems are:
- We get to see how the system would have performed had we taken every trade that technically signalled, regardless of news events. In other words, whereas we may avoid a trade now because of upcoming news, in backtesting we do not. This gives us a benchmark for live testing, and comparing the results we get from live testing and backtesting we can get an idea of just how much effect news items have on the system.
- We get a rough idea of the metrics of the system: win to loss ratio, drawdown, average win size and average loss size etc
- We can quickly discount a system if it performs poorly in the past, as it is unlikely to turn around its performance in the future
- Even though the results cannot be 100% relied upon, it is far, far quicker to get a feel for how the system might perform through back testing than it is through live testing going forward.
How to backtest a forex strategy with manual testing software:
Interesting New Kid on the Testing Block: