I recently received a questions via email from one of my readers about backtesting day trading strategies and the proper length of time that one should test before implementing them. I felt would be better addressed in a post so others could learn and hopefully contribute to the dialogue, any of their experiences regarding this matter. I’ll answer this based on my experiences as best I can.
Oddly, my questions don’t have much to do with analysis, it has more to do with psychology and readiness. I am an Optionetics student (Options trading education), and went to one of those “free seminars” about 2 years ago. They are good at teaching fundamentals, and even providing some trading ideas to get started.My questions are related to how much back and virtual testing is enough? I have 1 system, but it’s taking months to test it. Does this sound normal?
I’m almost always testing some sort of scan or strategy because I like that mad-scientist approach where I can express my creativity within this trading world. I can say that most of my indicators, chart patterns, and thought processes are far outside of the norm and I want it to stay that way to capitalize on being different than the crowd. I’ve been fortunate to have lived through the 98′ blow off market top and the 08′ meltdown so I’ve seen most of what the market can throw at us investors. This brings me to one of the first things you have to do when testing a pure market timing strategy
- Which side of the market are you playing, and more importantly, where has the market been lately relative to the longer term. You would never test a long strategy during a phase of market contraction. Also, your results would most definitely be different if your backtesting took place at the beginning of a bull market vs. the latter stages of one. This is one of the reasons I have an issue with backtesting different strategies without being aware of the type of markets, as the circumstances are never the same for comparisons sake therefore it’s quite difficult to trust backtesting results.
- Is the system one for the general markets or is it targeting a certain setup for individual long/short opportunities? These two types of trading are significantly different and require a different approach regarding the reaction time to enter the trade. It’s not as simple as finding a long strategy that works and then reverse engineering it for a shorting screen because in my experience that rarely, if ever works. When shorting you are capitalizing on fear and that behaves completely different than hope/optimism (long trade) and requires you to be quicker on the trade trigger and have a shorter trade expectancy.
- Develop, implement, and strictly follow mechanical rules for entries and exits for when you’re cutting losses and taking profits. Know exactly what you are going to do ahead of time so you’re not changing your approach mid way through. There is no room for being a discretionary trader when testing a strategy because then you’ll have nothing to test future results against.
- If you are testing a market strategy for going bullish/bearish I would give it 2 complete cycles of the markets being bullish and bearish, and that would be by whatever indicators you determine a bull/bear market to be identified as. If you are testing a bullish strategy for a certain chart pattern, I would give it at least 2 bullish cycles to test and make sure it works in different bullish up cycles. The same goes for bear strategies. This could take anywhere from 3-6 months depending on how volatile the markets are.
- Start with a small percentage of your capital when you feel you have done the appropriate amount of research and increase as you have success. It’s almost inevitable that your results with real money vs testing money are going to be different because you’ve introduced emotions into the mix, but be mindful of any variation in your trading and correct it as you go. This is a very important step because if you start off on a losing streak your first inclination may be to scrap the whole system you’ve worked on for months, but this is the reason why you start out with a small position size. Give your system time to play itself out. A core Turtles axiom regarding this is:
Judge yourself not by the outcome, but by your process.
This list is my no mean all inclusive of what goes into testing a trading strategy, but it’s a start. If after posting this I feel like there is something more I want to add I’ll revise it, or maybe a comment by one of my readers will spark my memory but there’s no point in holding this post because I may have forgotten something.
