Trade with Logik Episode 4
In episode 4 of the Trade with Logik podcast Matty and Shahan discuss the various problems with manual back testing your trading strategy. We cover the factors such as over-fitting, back test time lag and psychological issues in back testing. Give it a listen above!
If you don't have time to watch the full thing, don't worry I'll note down the key points in this article.
One of the main topics discussed in the podcast was Over-fitting. This term relates to traders testing their strategy on one asset class, then using that strategy on other asset classes. But when the strategy doesn't work as planned, the trader starts to tweak at it to make it fit the new data. What you get is over fitted results which are not true results of your strategy.
For example, throughout 2018-2019 you use a Support and resistance strategy with fixed criteria before entry. But then, you go back in time and start adding factors like candle formations, other indicators etc. Until the back test results are looking really good. What tends to happen is your results look amazing in back testing but... when it comes to applying it live, it's trash. This is because you've fitted the back test to suit whatever happened between 2018-2019 for that asset. Market conditions never stay the same, so fitting the results like that have terrible effects on live results.
As mentioned by Shahan to reduce and remove over fitting from your results, you test your strategy on one set of data known as test data. Then that's your strategy, you then apply it to other asset classes without changing the strategy at all.
2. Trading hours
One interesting point that was mentioned is the hours of the day you are able to trade. When you're back testing, you go through data that actually is released over a 24 hour basis. What traders don't realize is they start to record data they may not have actually caught during hours of the day they trade.
E.g. I can only trade between 5 -10pm. If I'm back testing, how many of the trades can I actually enter in that time period. That was one of the problems of many day traders and scalpers out there. Your back test results may not show what you can truly catch during those time periods.
So, to prevent this from happening make sure to filter out trades that may occur at times when you couldn't have entered the trades. However, if you're using an EA that you developed with your strategy you may not need to do this.
3. Psychological Biases
In back testing, you'll have periods where you're full focus. However, when it comes to it in reality you can't be fully focused 24 hours of the day, like you are in back testing. This is something to account for as ego depletion and fatigue starts to kick in if you're at your screen all day waiting for setups.
So, when you're back testing you'll never experience the live psychological issues that go on. E.g. if you happened to have a disaster in your life, your mindset and focus on trading may not be 100% which causes your decisions and trades to be off. So, when you find yourself not feeling yourself, tired or depressed/ euphoric you'll want to bring your mind back down to a neutral objective mindset.