The Pitfalls and Benefits of “Back-testing”
When I first started trading I would have spent hundreds of hours poring over charts back-testing my trading strategy.
I even enlisted my girlfriends help to record my back test results in a spreadsheet (much to her chagrin).
But what exactly is back-testing and does it help?
Two types of back-testing
There are two main types of back-testing, each with its own strengths and weaknesses.
Manual back-testing is when you sit down and scroll though historical charts looking for pervious occurrences of your trade set-up to see if they have an edge over the market.
For example if you have a trading system that buys when two moving averages cross with a 15 pip stop-loss and a 30 pip profit target, then you would go back in time in your charting package to see what happened on previous cross-overs. You then record the results of at least the previous 30 trades (always a minimum of 30 for statistical significance), and see if you would have been profitable if you had taken them.
This can be a helpful process to work out if your strategy is viable. In addition it can be a useful way of “getting to know” the market you are trading.
On the downside you need to be aware of your own bias to see only the positive results, and the impact of slippage and costs that you cannot see all that clearly from your manual back-test.
Rise of the robot
The second type is computerised back-testing. You have your strategy coded and you run it on historical chart data and then get a report based on the results.
Computerised back-testing eliminates mistakes of human bias, but it is a difficult science to master. You have to deal with:
- Mistakes in coding the system compared to your vision
- Problems with past data that may not be accurate
- Curve-fitting your strategy to fit past data
You also need a very good understanding of what reporting metrics you should be looking at.
Some great trading strategies cannot be back tested
It is particularly important to note that many good (and great) trading strategies cannot be back-tested.
If you limit yourself to strategies that can only be back-test, or only perform well in back-testing then you may close yourself off from reaching your full trading potential.
While the idea of having a perfectly back-tested and automated approach to trading may sound great, only about 10% of really good traders are automated traders, the rest are “rules-based” discretionary traders.
Your plan never survives contact with the enemy
One of the big challenges is that your system will very rarely perform like its back-test in real-life. Your performance will very much depend on the market type the strategy was designed for vs. the market type we are in now.
Be very wary of your back-tested system when you first trade it live, and notice when if it’s working as expected.
To back-test or not to back-test?
So should you back-test?
Sure, if you are aware of the pitfalls of back-testing and you use it as a tool to achieve certain goals.
Just don’t fall prey to idea that having a back-tested system is the be all and end all and forget to trade trade what is in-front of you.
If you have back-tested in the past, has it worked for you?
(P.s. thanks to Mary who suggested this post)
About the author
Sam Eder is a currency trader and author of the Definitive Guide to Developing a Winning Forex Trading System and the Advanced Forex Course for Smart Traders. He is a key team member at premium FX services provider www.fxww.com and part owner of Forex Signal Provider www.fxrenew.com (You can get a free trial). If you like Sam’s writing you can subscribe to his newsletter for free: https://fxrenew.com/newsletter-sign-up.