Crude Oil Technical Strategy: Trading the Relative Strength Index (RSI)

Read oil market news and you're bound to come across references to technical analysis, which is viewed by many traders as an important tool in predicting price changes and signalling market momentum. While most will say that technical analysis should be used as a confirmation of fundamental views, there are many traders who claim to have successfully implemented trading strategies based solely on quantitative and technical analysis. In the first of a series of posts, I'll be looking at whether a simple momentum indicator can be used to create a profitable oil trading strategy.

The first indicator we’ll look at is the Relative Strength Index (RSI). The RSI is a momentum oscillator; it cycles between a value of 1 and 100 and measures the trading-momentum of a security.  A higher measure indicates a security could be overbought and a lower number oversold. (For more on the calculation of the RSI, check out this article at Investopedia.) A rough guide is that anything over 70 represents the security being overbought, and anything below 30 represents oversold. A potential strategy therefore can be to go short when the reading is over 70 and go long when its below 30. While there are plenty of articles out there discussing the strategy for stocks and currencies, how does the strategy play out for crude?

To take a first-look at the strategy, I’ve downloaded a price-set from EIA and created a 14-day RSI for the beginning of 1986 to the beginning of 2013. Using a VBA-script, the file loops through the data and works out the return from a buy and sell strategy in which the upper and lower RSI limits vary, thereby allowing for values of the upper and lower bounds to vary around the 30 and 70 marks. 

The table below shows the results. The non-strategy return, that is buying crude on day 1 and selling on the last day, results in a return of 389%; an equivalent annual rate of 6.1%. As the highlighted cells show, the best strategies are located around the (30-40, 76) limits or the (40-44, 60-64) limits, with the two best results being (44, 62) and (40, 76).

When using the (40,76) options for the lower and upper limit, there are only 88 days when the RSI is over 76 but 1208 when it is below 40, out of a total 6828 days in the data set; thus the strategy is almost entirely just predicting a few key selling points. In contrast, the (44, 62) presents a more-even strategy as there are 1829 buys and 1259 sells. (44, 62) however is a long way from the orthodox 30 and 70 mark; so is RSI really applicable in this case, or are we just getting lucky?

The truth is, the RSI alone is probably not going to give us a profitable trading strategy. As OptionAlpha points out, a more powerful indicator can actually be an RSI divergence. This is when the price reaches a higher high (lower low) but the RSI fails to reaches below its previous high (low). Graphically we mean something like below (taken from OptionALpha). This could explain why less extreme RSI thresholds are actually giving a better return. Alternatively, it may be we need to combine the RSI with another indicator to give a better indication of possible trends. In fact it’s common to use the RSI with a moving-average indicator; something we'll look at in more detail in the next post in this series.