I have written posts in the past about my progress in developing and testing new trading systems so I thought it might be helpful to start more at the beginning. Here are my steps for developing a succesful (hopefully) trading system.
1. Identify My Edge – my trading edge is the reason I believe I can make money in the market. I am a technical trader, so my edge consists of the way I “measure” the market via patterns and indicators to determine profitable trends, entry points and exits.
2. Define My System Rules – here I take my edge which hopefully was well described in good detail in step 1, and write down the specific rules I am going to follow when trading. This step needs to be very specific…when will I trade, how much will I trade, when will I exit, etc.
3. Evaluate My System – once I have my rules well defined it’s time to start trading…in two possible ways. First, I could paper trade. I recommend paper trading for a minimum of 6 months just to get a feel for how the system is really going to work. The second way to evaluate my system is via historical data and back-testing. This is the way I prefer to trade and I developed my systems using 10 years of historical data.
4. Revise My System – based on the evaluation I will likely find reasons to revise the system. Even when I am the most careful and thorough in designing a system, after going over my back-tested trades it always surprised me the one or two ways the rules actually ended up being implemented. I use Amibroker for my system testing and I program my system on charts so I can see the actual entries and exits that are triggered. Without seeing these results, it’s hard to really understand the various trading circumstances I am be presented with.
5. Start Trading – This is where the “rubber hits the road.” All the preparation is now over and the fun and terror begins. Real trades with real money. I like to break my trading of a new system into 20 or 40 -trade “test baskets”. I do this as much for psychological reasons as for any valid statistical reason. I tell myself that I won’t get excited or dissappointed if the system does not perform as I expected until at least one full test is run. Usually, I will follow up the first test run immediately with another to see further validity of the system.
6. Deal with the Psychological Part of Trading – this part of designing a system actually is the most important and the most difficult. It actually takes place along with step-5…because I am trading now and have to deal with the emotions of the trades. I don’t know why it is…or why it seems like it…but I swear my system implementations always start out going really well, much better than expected, and then I get into the muck of numerous losing trades, drawing down most of my initial profits. I’ve gotten used to it so I can better deal with it, but it’s still annoying. I find that keeping a trading journal or trading thoughts document helps. Talking with other traders also helps. Because the psychological issues of trading are not shared by many other vocations it’s hard for others to really understand what I’m going through.
7. Evaluate the System – I think this step in system design is very important but often overlooked. Just like other businesses, trading is a business that can be continually improved upon. I have to be carefule here and only evaluate and improve on a structured timetable. It’s important that my systems are given enough time to become statistically valid before I review results and consider improvements. This is hard especially during drawdown periods, but I really believe it is important to continue trading a system unchanged for long perdiods of time. 100 trades is much more indicative of a system than 20. If I change my system after 20 trades I really won’t know if changing it is for a valid reason.
After completing step 7, I find that it’s best to go back to step 1 again, albeit more briefly, and do the steps again. As a trader I’m constantly trying to improve my craft, but a structured, thoughtful approach will result in greater success…