From the November 01, 2011 issue of Futures Magazine • Subscribe!

Clearing up price level trading

Strategy backtesting

Because simple visual inspection can be misleading in determining the efficacy of a trade strategy, technical traders are encouraged to backtest. Again, modern trading platforms make it easy to develop strategy ideas and conduct backtests on historical data quickly and without specialized programming knowledge.

"Ups and downs" (below) shows an equity curve of a three-year backtest of the MACD, money flow and RSI indicator strategy. A strategy that makes use of the indicator combination for entry, and employing an initial four-point stop loss (moderate for the ES contract) and trailing stop position management is easy to write and execute on historical data. The backtest equity curve illustrates the importance of historical testing as well as some of the pitfalls of strategy backtesting in general.

For example, if a backtest period consisting of the last 200 trades had been used — the circled part of the equity curve — the trader might have been lured into believing he had found a highly successful strategy.

Strategy development addresses the mixed results shown by advocating the use of in- and out-of-sample data to minimize curve fitting and generate more accurate hypothetical results. Veteran strategy developers advocate lengthy periods of historical backtest, far exceeding a three-year test. The well-known systems technical strategist William Eckhardt describes backtest periods lasting decades before becoming convinced of any strategy (see "William Eckhardt: The man who launched a thousand systems," March 2011). The depicted equity curve illustrates the inherent difficulty of verifying new trade strategies.

Comments
comments powered by Disqus