(Click on Chart to see clearer view)
I wrote an article recently entitled "
Buying Weakness and Selling Strength" which detailed a trading strategy using the 2-period RSI as a buying (2 closes below 15) and a selling (close above 80) trigger. I've been using the strategy to enter and exit the TSM picks. By the tone of your questions, I can see that you're wondering about the use of stop losses to reduce risk. Bear in mind that the RSI exit is a dynamic trigger, that is, the only way to exit the TSM trade (profit or loss) is at the next day's open following a RSI(2)>80 close. That's not to say that you might use other criteria. Here, I want to briefly cite some pertinent statistics. Note though, that this dynamic exit can create a few large losses (see the plus 10 point loss at the extreme left of Chart I).
2008 was a bad year for the markets. The S&P, for instance, lost 38.5% of its value. As I cite in the article, the above RSI strategy produced 80% winners and a 1.6 average point gain (per share, per trade) in 2008. Chart I shows the distribution of points gained and lost for these trades. Note, rules defining the trade strategy can be found in the article. While the strategy is sound, you should understand its draw down characteristics and the idea of controlling risk by position sizing before following it in your own account.
Before going into those subjects, let me say that I view TSM's obligation to you is solely to identify quality TSM trades, i.e., what's ripe to trade. You have to choose the manner of trading that suits you. It's up to you to choose to use hard stops, trailing stops or some other manner of risk control. I will suggest TSM entry and exit criteria based on the RSI criteria and trade my account that way. Of course, this is subject to change in the future, and if that happens, I'll let you know what I'm thinking.
Back to the 2008 data above and some important statistics: average draw down was 6.7% during the life of the average trade; the average trade life was 6.6 days; the win rate was 80.6%; the average loss was 2.43 points; and the average win was 2.49 points. You can expect to win:
Expected Win ($) = (% Win)*(Avg Win) - (% Loss)*(Avg Loss) = (0.806)*(2.49)-(0.194)*(2.43) = +1.54 points per share per trade
Great strategy, but let's now consider the draw-down risk. If I were to use a static stop loss (say IBD's 7%), this strategy's return would be far less (nearly always true with strategies utilizing static stop losses). Instead, let's consider position sizing. Say you want to put 30% of your portfolio into this strategy (whether your portfolio is $10k, $100k or $1 million). You have the option of building a single position or dividing my 30% into multiple positions, for purposes of illustration here, say 10 positions. Consider the draw-down risk of the strategy for the entire portfolio, that's the important number.
Chart II shows how portfolio risk changes with the number of positions comprising the portfolio. The red line is the 1.54 point average from 1 to 10 positions per portfolio; the blue lines are the 1%/99% confidence limits; and the green lines are the 5%/95% confidence lines. For example, if I divided the 30% of my account into 2 TSM trades, I could expect a 7.5 point loss (average for the 2 individual trades) or greater 1 time in 100. Conversely, I could expect an average 12 point or greater gain 1 time in 100, too. Now, if I divide the 30% of my account into groups of 10 TSM trades, I could expect a far smaller average loss ($1 or greater loss 1 time in 100 for the group of 10), but also, a much smaller average gain ($4.1 or greater gain 1 time in 100). There it is, good trading strategy utilizing position sizing to minimize risk.
Other ideas that I plan to pull together in a more formal report: selling Puts on TSM stocks meeting the above TSM criteria; generating additional income with a Covered Call strategy for these type stocks; scaling into the stock positions with 2 or 3 entries rather than a single following 2 days with the RSI close below 15. I also plan to develop an Excel based program that will enable one to generate portfolio risk charts like Chart II utilizing a Monte Carlo approach coupled with a group of test trades that define the expected returns from your strategies.
Example of Controlling Risk by Position SizingAs an example, assume a $100,000 account that devotes 30% ($30,000) to trading TSM stocks. For calculation purposes, assume, too, that the stocks being traded are $40 stocks. Trader one puts his money in 375 shares of two stocks ($15,000 in each). On average, he can expect to make $1,155 (2 x 1.54 x 375) or 3.85% on his $30,000 every 6.5 days (if a new group of stocks are ready to trade); however, once in 100 trades of these type pairs, he can expect to lose $5,625 (2 x 7.5 x 375) or a 18.75% loss--perhaps on the very first trade made. Too, once in 100 trades he can expect to profit by $12 points to win $9,000 or a 30% gain. Personally, I want less variation in my returns.
Contrast the 2-position approach with a 10-position approach. Trader two puts her money in 75 shares of ten stocks ($3,000 in each). Again, on average, she can expect to make the same $1,155 (10 x 1.54 x 75) on her $30,000 every 6.5 days (again, if new trades are available). Now, once in 100 trades of these 10-position trades, she can expect to lose $750 (10 x 1 x 75) or 2.5% or make $3,075 (10 x 4.1 x 75) or 10.25%. This trading strategy (using position sizing to mitigate risk) experiences far less downside and upside variation.
My trading platform (Tradestation) costs me $1 per 100 shares or option contract in commission so the difference in the two approaches describe above is small. If you're using a broker that charges a fee by the trade, your commissions would be substantially larger. Consider broker ThinkorSwim for a commission structure similar to mine with even a $10,000 account, though you might have to ask for it.
If you have any questions, don't hesitate to ask.
Ric Miller, Ph.D.
6-Sigma, Master Black Belt
