The RSI & Price Trends by Kevin Luo
Tweak It, Test It
The relative strength index (RSI) is well known among technical analysts. As with all indicators, it tends to work better on some stocks or at certain times. Can it be modified to give better results? We’ll find out.
The relative strength index (RSI) is a popular technical indicator designed to measure the rate of change in price movements. All RSI values fall between zero and 100. One of the common interpretations of RSI is that the stock is oversold when the RSI falls below 30 and overbought when the RSI rises above 70. The initial impression of its application is that you buy when the stock is oversold and sell when the stock is overbought, or go short when the stock is overbought and cover when it is oversold.
HOW WELL DOES IT WORK?
To find out how well it works, I conducted a large-scale backtest on 1,816 random stocks traded on the NYSE and NASDAQ. The test data covered a 10-year period from January 1, 2005 to August 31, 2014. Given the scale of the backtest, the output is considered sufficient, precise, and reliable enough to achieve significant conclusions about the application of the RSI strategy. I applied a custom-built automated system—I’ll refer to it as the software—to run the performance tests.