Consulting | Custom Trading Systems
Built-In Indicators | Custom Indicators

May 2007 Stocks and Commodities Traders Tips

Moving Averages: Long on Talk, Short on Action

Original article by Anthony Trongone, Ph.D
AIQ code by Richard Denning

The AIQ code for Anthony Trongone’s TMA QQQQ trading system is given below.

The author shows how the addition of two filters to a moving average trading system can make a big improvement over the traditional way moving averages are used in trading systems. Dr. Trongone acknowledges that once the filters are added, the number of trading signals is reduced significantly.

Using the system on just the QQQQ does not utilize the capital sufficiently to give a satisfactory return on capital. One of the filters on the QQQQ system requires a fixed volume cutoff of greater than 100 million shares. I try to avoid using fixed cutoffs whenever possible because they are not adaptive to changing market conditions. Also I wanted to see if a trading system could be developed that would show a more acceptable return on capital. To do that I needed a flexible filter for the volume that would automatically adjust to different ETF trading volume and also to changes in the general level of market volume.

Thus, in place of Dr. Trongone’s volume filter, I substituted the rule that the prior day volume must be 110% higher than the 50 day simple average of volume. To generate more signals I needed to run the trading system on more than one ETF. The ETFs need to have sufficient liquidity to withstand a higher frequency of trading without excessive slippage. I scanned for ETFs that trade 1,000,000 shares over the most recent 50 day period that also had data back to or before 10/18/02, the beginning of the test period. There are 23 ETFs that passed this test: DIA, EWH, EWJ, EWM, EWS, EWW, IWD, IWF, IWM, IWN, IWO, MDY, QQQQ, SPY, XLB, XLE, XLF, XLI, XLK, XLP, XLU, XLV, and XLY.

I ran a long-only trading simulation using the AIQ Portfolio Simulator and the above list of ETFs over the test period, 12/31/02 to 03/12/07. I put 100% of capital into each trade. When there was more than one signal to trade, I choose the one with the highest 30-day relative strength. All entries are at the opening price and all exits are at the closing price on the day of entry (no positions are held overnight). Commission of $0.005 per share was deducted. I did not factor in any slippage. The results of this test were favorable-see Figure 1.

FIGURE 1 (Click here to view).

Figure 1: Equity curve for the modified TMA system versus the NDX index using a list of 23 liquid ETFs with more than 5 years of trading history.

I also ran a simulation over the generally bearish market period, 12/31/99 to 12/31/02. The results of this period showed an average annual loss of 2.61% and a maximum draw down of 23%. This compares favorably to the NASDAQ 100 which showed a 24.4% annual loss with a greater than 80% maximum draw down.

EDS Code For Tronegone Moving Averages:
Moving Averages.EDS




















Disclaimer | Privacy Policy | Contact Us | Spam Blocker | ©2009 Traders Edge Systems