I initiate a journal here as my New Year resolution to discuss one of my trading strategies. The object of this strategy is to generate 15~20% annual return by swing trading US stocks without too much stress. The strategy is based on EOD data and generates entry signals at the end of day. We enter our positions throughout next day and exit our positions after 8~11 trading days. Personally I favor to hold the positions for 11 trading days. Currently this system is purely based on price movement; it does not use any fundamental information. It will be very interesting to see how much improvement we can get if we add the fundamental information into the system. I invite your comments and suggestions on the systems. Please keep in mind that this is a journal to discuss a virtual trading strategy, I may or many not own any positions mentioned here. You trade at your own risk. More discussions will be followed. Jim
Simulated performance of the system Year Annual Return% 1990 32.8 1991 27.4 1992 16.9 1993 11.4 1994 6.2 1995 17.5 1996 23.3 1997 1.5 1998 56.0 1999 54.8 2000 38.9 2001 36.6 2002 17.9 2003 25.7 2004 14.4 2005 10.6 Avg% 24.5 Sharp Ratio 1.53 The simulation was based on the end of day price. 0.6% per trade was deducted in the calculation to account for commissions and slippages. The max number of stocks to enter is 10 stocks per day.
A few more issues about the system: 1. Monthly max drawdown: -14.1%. 2. Stop loss: Not implemented. 3. Position size: 2% of the total capital at most. 4. Margin: Not used in the calculation. The system is initially designed for large capital capacity. It is intended to generate consistent positive returns over the years. I have been using discount brokers to pay $3 or less per trade. The system will generate 1000+ trades per year. The commission really makes difference in the end of the trading game.
Thanks for the question. Without talking too much details of the system, the general ideas of the entry/exit criteria are: Entry: Buy long when stocks are oversold, i.e., nobody wants to have them. Sell short when stocks are overbought. The final selection is dependent on market condition, price movement, statistical probability and other factors. Exit: A fixed 11 trading days holding period. Please note that it is not 11 calendar days. The period of 11 days is selected based on overall yearly performance and capital capacity. Jim
Here are the long entries on 1/11/06 Symbol Share Price TRCA 1600 7.21 FRNT 1500 7.89 JBLU 800 14.06 KOSP 200 49.34 LIOX 1700 6.76 PGIC 1400 8.26 SIL 800 14.61
One can see that those you have listed do indeed represent a fair degree of bearishness/oversold conditions, but what are you using to determine the potential for a reversal? FRNT, for example, has the following against it: 1) 7 down days on increased volume each day 2) Recently crossed under its 10-day MA 3) Today was a *huge* down day with a high just barely above the open 4) Stochastics, OBV, MACD, etc. and even ADX all indicate continuation of the immediate trend. Please understand sir I am not attempting to challenge your choices or position; rather, I seek clarification as I would hope to potentially learn something. Kind regards.
Thanks for the comments. I design/backtest many systems based on common indicators such as MA, Stochastics, OBV, MACD, ADX etc, they all did not work out well in terms of annual average return and Sharp ratio. The system I introduced here uses a very different approach. It is mainly based on probability, rather than those common indicators. It is a buy-low and average down system with some stop measures in place (only 11 holding days). The stocks picked by the system look very ugly in charts, that is why it makes money in the simulation. The edge of the system is: buy long when nobody wants to do so and vise versa for shorts.
Sounds risky to me. Will you be adding to your position if tha target stocks continue to fall in price? How do you determin position size? I see all your picks are at EOD price. Is this when you enter the trade? Best of luck.