Predictive Correlation

By: Murray A. Ruggiero

The following is an excerpt from Murray A. Ruggiero's Cybernetic Trading Strategies

A correlation between two markets does not always mean that the current movement in one market can be used to predict the movement in the other.   To address this issue, I have developed a concept called predictive correlation. The idea behind predictive correlation requires taking a correlation between an indicator N periods ago and a change in a given market over the last N periods.  For example, on daily data, we can take a correlation between T-Bonds[5]– T-Bonds[10] and the S&P500-S&P500[5].  This correlation will tell us how predictive a simple momentum of T-Bonds has been over the length of the correlation.  The predictive correlation curve is much different from the curve generated by standard correlation, but it does seem that they both trend in the same direction.  The power of predictive correlation is that we can correlate an indicator or inter-market market relationship to future price movements in the market we are trying to trade.  This allows us to use relationships and indicators in rules, and to trade these rules only when these indicators are currently predictive.  Let’s now add predictive correlation to our modified S&P500 pattern.

We use Close[1] of T-Bonds-Open[1] of T-Bonds as our independent variable, and Close-Open of the S&P500 as our dependent variable.  We go long on Mondays when a 35-day predictive correlation is above 0.  The amazing results, from 4/21/82 to 7/26/96, are shown in Table 8.6. 

This system produced over $600.00 per trade, after deducting $50.00 for slippage and commissions.  We won 66 percent of our trades and had a profit factor of 3.75.  These numbers are much better than any of the variations that did not use predictive correlation, and they should prove the power of predictive correlation.

Results of Adding Predictive Correlation

Using the CRB and Predictive Correlation to Predict Gold

One of the most powerful methods is a ratio between the inter-market and the commodity being traded.  I will now show you how to combine the ratio between the CRB and gold with predictive correlation to develop a very profitable and reliable system for trading gold.

The Commodity Research Bureau index (the CRB) is a basket of 21 commodities.  This index has been traded as a futures contract since mid-1986.  It has had an inverse correlation to T-Bond prices and it has been positively correlated to gold during its history.

On the basis of this relationship, I decided to use the ratio between the CRB and gold to develop a trading system for gold.  When this ratio is moving up, the CRB is outperforming gold, and gold should catch up.

Another fact about gold was revealed in my research.  Often, when the CRB starts moving gold will first move in the opposite direction and test support, before moving in the direction of the CRB.

On the basis of my understand of the gold market, I am proposing a system that (1) uses a moving-average crossover of the ratio between the CRB and gold to generate its signals and (2) enters the market on a limit order set at the level of N-day exponential moving average.  This concept was tested on back-adjusted contracts over the period from 11/18/86 to 7/26/96.  The system is based on a sound premise.  If inflation increases, so will the price of gold.  Still, it performed badly and made only $4,000.00 in profit over the test period.  The reason the system did so badly is that it had large drawdown during periods when the CRB and gold decoupled.  We can filter these periods out by using correlation analysis.  Let’s now add the predictive correlation between the ratio of the CRB/gold 5 days ago and the past 5-day change in gold.  This simple gold model, coded in TradeStation’s Easy Language with parameters selected based on my research, is shown in Table 8.7.

Gold/CRB Ratio System

RACorrel is a user function developed by Ruggiero Associates.  It calculates the standard Pearson’s correlation found in any statistics textbook.    Table 8.7

The model made over $48,000.00 during this period, and the system was profitable on both the long and short sides.  Another important point: The entry method (buy on a limit set at the 80-day exponential moving average of the close) increased the average trade by over $500.00 when compared to the method of entering at the next open when the signal first occurs.

The system does have some problems, however.  For example, the average winning trade lasted 45 days but the average losing trade lasted 144 days.  We can help solve this problem by developing better exits for the model.  Even with this problem, the model is fundamentally sound and could be the core for a system for trading gold futures or gold mutual funds.

Gold/CRB Ratio System Results