An Overview of Backtesting and Paper Trading
As the sole owner and operator of a trading business, you play the role of everyone from CFO to data entry clerk. You are also the entire QA department. And like any company, if you want to be assured of the quality of your product, you need to test it.
Unless you are detail-oriented, you may not particularly enjoy the role of tester. But testing is a crucial task to perform if you want to give yourself (and make no mistake, you are your own most important client) a successful swing trading system.
Nevertheless, testing is a long and tedious process. It will try your patience. You will be in the midst of testing, see positive results, and think: 'If only I were trading for real, I'd be making money right now. I should end these tests and get on with it already!' To be sure, the siren call of live trading is strongest right after big profits are made on paper.
This is a temptation to resist at all costs. Do not cut your testing short. There is simply no other activity that will identify problems with your trading system as effectively as running backtests and making paper trades.
During testing, for instance, you may find you are taking on too much risk and need to adjust your trading plan to reflect the harsh realities of the market. Or you may discover that your intended basket of resources do not work well together and is in need of compatible alternatives.
The component that most directly benefits from testing, however, is the trading strategy. Often times testing reveals how a strategy requires much more than a slight tweak to make it work. Other times even drastic overhauls will not save it. At which point you will have to abandon it outright and start again with a clean slate. This is no doubt discouraging. But take comfort in the fact that you may have just saved yourself from financial disaster.
There are qualitative benefits to testing as well. These are just as important as those which quantitatively refine the profit potential of your trading system. Testing, for instance, is a great way to learn the ins, outs, and duration of trading. Think of it as a free education where you simultaneously gain valuable hands-on experience.
At a deeper level, testing poses a basic question: Do you have what it takes to be a swing trader? If you are interested in finding out sooner rather than later, make sure your testing procedures mirror the way you will actually conduct your trades. Then if you find you do not care much for trading after all, this life lesson comes virtually free of charge.
But if you can discover the joys of swing trading even in the testing phase, it will be much easier to keep your emotions in check and resist the lure of live trading before you are ready. After having spent so much time developing your trading plan, trading strategy, and selecting your trading resources, you owe it to yourself to thoroughly test how these components work together. Doing so greatly increases the chances that your swing trading system will be profitable once you expose it to the markets.
Backtesting is time consuming, especially for the discretionary trader. Many backtesting sites have sprung up in recent years to address this problem. Unfortunately, most of them only allow the tester to work with a select set of testing parameters. While some permit customization of their set, others do not. But even when customization is possible, it is unlikely any one site will have everything you need. Worse still, there are things in the trader's universe that are not quantifiable at all. Take trend lines, for example. Although much can be done to standardize the way they are drawn, they nevertheless remain more of an art than an exact science. So even though you may have programmed very little wriggle room into your trading strategy, automated backtesting sites may not be much use to you. This is doubly true if your trading strategy is sufficiently complex so as to defy automation altogether.
What you should not do is construct a swing trading system around what these sites are capable of testing. Why put yourself at the mercy of site programmers when you do not have to? This defeats the purpose of doing it yourself and generally dampens the entrepreneurial spirit.
There is a better way. Consider making backtesting an offshoot of paper trading. There are many ways this can be done. But do not skip backtesting altogether. Why? Because you want some idea how your trading system will behave in a variety of market conditions, and not just in the one that currently exists today. At minimum, you want to see if your system works in both bull and bear markets. But what about in volatile and non-trending markets? Or during seasonal and cyclical market patterns? Select some well-defined periods of the past, run your tests, and find out.
Paper trading is live trading minus the actual risk. No real money is at stake. You execute trades using a fake brokerage account. Your wins and losses are recorded 'on paper' only.
But does this mean you should go hog wild and trade whenever the mood strikes you? Absolutely not. You must trade exactly as you intend to trade when you go live with a real account. Or as close to it as possible. Otherwise, you will not be able to tell if your trading system works.
This means paper trade in real-time. You might be tempted to wait until the end of the day or week, and then test retroactively. But it is important to get a feel for the rhythm of trading. Paper trading not only tests and refines the material components of your trading system. It does the same for your skills as a swing trader.
One reason why paper trading might be preferred to (but not a substitute for) backtesting is because of its focus on the present. Testing in today's market should have greater predictive value for your trading system tomorrow than if all testing weight was given over to the distant past. This is important because you will soon be live trading. The rationale here is similar to what drives some traders to prefer weighted and exponential moving averages over the simpler, arithmetic version.
Just be sure to note how many trades you are making. On the one hand, it is a problem if your system generates too few trades. You do not want the funds in your account to idle in cash for long, but to constantly turn over from one trade to the next. On the other hand, if your system generates more trades than your real account can handle, trade them all anyway on paper. For this is a great opportunity. Not only will you quickly reach the magic number needed to make your testing statistically significant (30+ trades minimum, 100+ better), you now have the luxury of being choosy. Identify which trades are the best and what makes them the best. Include those characteristics as part of your trading strategy's selection criteria for future trades. Or else have them be criteria that 'kick in' whenever the number of potential trades exceeds the limits of your trading account.