Now I will present the system to be used for the study that was described in Part One.
The system uses an entry and exit signal generated from the CCI (Commodity Channel Index) indicator. For more information on the CCI, here is a helpful explanation: Commodity Channel Index.
As systems go, this is not one of my favorites, although it is profitable. I have not spent an inordinate amount of time developing my version of it. As the system makes money from being contrarian, it will work well for this study.
All entry and exit signals are based on the end-of-day CCI(8) reading and will be triggered on the next market open.
A Buy signal is generated when CCI(8) crosses above -150 from below. This means that the system is buying weakness.
A Sell Short signal is generated when CCI(8) crosses below 90 from above. This indicates that the system is shorting on strength.
An exit signal for any long position is generated when CCI(8) is rising and crosses above 40 from below, meaning the system is selling the long position into strength.
A Buy to Cover signal for any short position is generated when CCI(8) falls to below 50, meaning the system is covering into weakness.
System Specifics and Results
Period Tested: 10 years back ending on 12/31/2008 using the SPY.
10,000 per trade with no compounding of gains. Gains are purposely not compounded during back testing. It will be best to cover the rationale for not compounding gains during proof-of-concept testing in a future post.
Commissions of .01/share are included. I have not included any returns generated from the cash available when the system is out of the market (It is only in the market about 50% of the time).
Some Chart Porn
The equity curve is fairly smooth, except for the end of 2008 (This will be improved in Part Three).
The weekly drawdowns are not awful, but will also be improved in Part Three.
The above graph shows the recent trades and the CCI indicator in the lower pane. Also plotted is the 50 day moving average. The 50 day moving average will be the intermediate trend indicator. Note the long entry (LE) that was made on 10/1/08. The CCI system entered long in oversold conditions, and then the market cratered, getting more and more oversold. This particular stretch of market history caused a lot of angst for traders of mean-reversion systems.
Presented above is a mean-reversion (contrarian) system. As such, it seeks to buy when the market is going down and sell when it is going up. Unfortunately, as shown in the October trade, the system will often take a trade against the intermediate trend (as shown by the declining 50 day moving average). Occasionally these trades against the trend will result in large drawdowns.
Now that the system has been presented, Part Three will use the intermediate trend (50 day average) to determine position-sizing and leverage. The final result will be the proof-of-concept of a “best of both worlds” system with decreased drawdowns and increased profits.