Those printouts provide a false sense of security. You may have a very nice printout, but what if the system delivers five losses in a row, while you trade real money? Nothing in your testing has prepared you for that, but it happens all the time. You grit your teeth and put on another trade. Another loss. Your drawdown is getting deeper. Will you put on the next trade? Suddenly, an impressive printout looks like a very thin reed on which to hang your future, while your account is being whittled away.
The attraction of electronic testing is such that there is now a small cottage industry of programmers who test systems for a fee. Some traders spend months, if not years, learning to use testing software. A loser who cannot admit he’s afraid to trade has a wonderful excuse that he is learning new software. He’s like a swimmer who is afraid of water and keeps himself busy ironing his swimsuit.
Only one kind of system testing makes sense. It is slow, it is timeconsuming, and it does not lend itself to testing a hundred markets at once, but it’s the only method that prepares you for trading. It consists of going through historical data one day at a time, scrupulously writing down your trading signals for the day ahead, then clicking your chart forward and recording the trades and signals for the next day.
Begin by downloading your stock or futures data for a minimum of two years. Swing to the left side of the file, without looking at what happened next. Open your technical analysis program and a spreadsheet. The two most important keys for traders on a computer are Alt and Tab because they let you switch between two programs. Open two windows in your analytic program—one for your long-term chart with its indicators, the other for the short-term chart. Open a spreadsheet, write down your system’s rules at the top of the page and create columns for the date, entry date and price, and the exit date and price.
Turn to the weekly chart and note its signal, if any. If it gives you a buy or sell signal, go to the daily chart ending on the same date to see whether it gives you a buy or a sell signal as well. If it does, record the order you have to place in your spreadsheet. Now return to the daily chart and click one day forward. See whether your buy or sell order was triggered. If so, return to the spreadsheet and record the result. Track your trade day by day, calculating stops and deciding where to take profits.
Follow this process throughout your entire data file, advancing a week at a time on the weekly chart, a day at a time on the daily chart. At every click write down your system’s signals and your actions.
As you click forward, one day at a time, the market history will slowly unfold and challenge you. You click and a buy signal comes into view. Will you take it? Record your decision in a spreadsheet. Will you take profits at a set target, on a sell signal, or on the basis of price action? You are doing much more than testing a set of rigid rules. Moving ahead day by day, you develop your decision-making skills. This one-bar-at-a-time forward testing is vastly superior to what you get from backtesting software.
How will you deal with gap openings when prices open above your buy level or below your stop? What about limit moves in futures? Should the system be adjusted, changed, or scrapped? Clicking forward one day at a time gets you as close to the real experience of trading as you can ever get without putting on a trade. It puts you in touch with the raw edge of the market, which you can never experience through an orderly printout from a professional system tester.
Manual testing will improve your ability to think, recognize events, and act in the foggy environment of the market. Your trading plans must include certain absolute rules, most of them concerning money management. As long as you stay within those rules, you have much freedom in trading the markets. Your growing levels of knowledge, maturity, judgment, and skill are much more important assets than any computerized testing.
Read More : SYSTEM TESTING
