 |
September 1999 Letters To The Editor
WEBSITES FOR TRADERS
Editor,Your new Websites For Traders section is an excellent addition to your magazine. As a technical analyst I am always on the lookout for new and interesting Websites devoted to technical analysis.
R.L. Robke, via E-mail
Dayton, OH
CHARTS & GRAPHICS
Editor,I read and enjoy your magazine each month with great enthusiasm. I have learned a lot from it and will continue to do so. There is nothing else that compares. My complaint is that the size of the charts is way too small. This is technical information with a lot of detail. Can you make the charts and text larger? I'll bet you'll get many compliments.
Joe Gnandt, via E-mail
Our intent is certainly to make Technical Analysis of STOCKS & COMMODITIES, The Traders' Magazine, as easy to read as possible. At the same time we want to make sure that each magazine issue has a variety of articles that cover different trading techniques. The tradeoff has always been the number of good trading articles per issue versus article length and chart size. In upcoming issues, we will publish some concise articles about the use of individual chart formations and computerized trading methods. These articles will focus on describing how to recognize, calculate, and apply an indicator to trading the markets. We will continue to use many different vendors' computer software to produce chart examples in our articles. At the same time, we will experiment with using larger charts and more printed page area. You can look for these changes, and thanks for your feedback. -- Publisher
STOCHASTICS EXPLAINED
Editor,Please explain stochastics. When I look at charts on the Internet with stochastic indicators plotted, no two versions of the stochastics indicator seem to be the same. Is there a generally accepted formula for the computation of %K and %D?
T. Rodgers, via E-mail
See my article on stochastics in this issue. Thanks for your interest.-- Stu Evens, Staff Writer
ALLOWING FOR INFLATION
Editor,I became a subscriber to S&C last year, and I like reading it every month. It is useful and full of valuable information. About the May 1999 article "Using A Constant Investment Size For Stock Trading Systems" by Jack Schwager, I have the following remarks. I understand and agree totally with Schwager's conclusion in the article, but since we must test a system over 10 to 20 years of data, don't you think it's necessary to also calculate the inflation rate? For example, Schwager uses $1,000 in 1985 and the same $1,000 in 1999, which would be worth less in "real money" if you take inflation into account. I suggest Schwager's test results would be even more realistic if we included an inflation factor. If we assume an inflation rate of 2.5%, this means that our $1,000 in 1985 should be worth about $1,413 in 1999. Is my thinking valid? I look forward to your comments.
Jack Schwager replies:Guy Hacha, via E-mail, Belgium
Mr. Hacha raises a good point, one that requires some reflection. I think the answer depends on the assumption. If we assume an investor will invest the same nominal amount over time, despite the fact that real dollar value is decreasing due to inflation, then using an inflation-adjusted factor would be a more accurate approach. However, if you argue that a $10 gain in a stock in 1999 is worth less than a $10 gain in 1979 because of inflation, which is certainly true, I think it is then also reasonable to assume that, all else being equal, the investor would have invested a smaller nominal dollar amount in 1979 than in 1999 for the same reason. Therefore, I think it's reasonable to make the simplifying assumption that the investor's nominal dollar investment size will grow over time to reflect inflation, or equivalently, that the investment size will remain constant in inflation-adjusted dollars. This certainly makes more sense than assuming that the investment size will remain constant in nominal dollars.
If the investment size remains constant in inflation-adjusted dollars, the two effects will cancel each other out. For example, if inflation reduces the value of the dollar to one-third of its value during the starting year, each $1 gain in a trade will only be worth one-third as much as a $1 gain in the starting year. However, for the same reason, the size of the investment can be assumed to be three times as large in the latter year.
TIME SERIES ANALYSIS
Editor,Michael P. Turner's July 1999 article "Correlation Among Stocks" presents wrong, incomplete, and/or misleading information regarding the use of correlation coefficients in the analysis of time series data.
First, two time series may be correlated simply because they are moving in the same general direction, that is, up or down, not because movements in the series themselves are correlated. High correlations between time series -- Turner mentions 0.9 -- almost always reflect trend when present. If present, the trend must be removed first. Perhaps the simplest method to remove the trend is taking the first differences for each series (subtracting yesterday's value from today's value, for example) and then and only then calculating the correlation between the series.
Second, the article implies the often incorrect assumption that the strongest relationship between two series occurs in what might be called the statistical present time. However, it may also be the case that the strongest statistical relationship entails some sort of lead/lag relationship, a topic unfortunately omitted from Turner's article. For example, the current or closing price of one stock may be a good predictor of the price of a second stock one or two hours or days later. These relationships can be explored using the same spreadsheet techniques suggested by Turner. One need only adjust the time relationship between the series, that is, correlate stock A today with stock B yesterday, two days ago, three days ago, and so on. In each instance, calculate the correlation only after differencing each series (with itself).
Third, it is simply not true that if two series are perfectly correlated, then a one-point increase in stock A would necessarily be matched by a one-point increase in stock B. This assertion would be true if and only if: 1) both series were measured in the same units (for example, dollars), 2) both series had the same mean (average over the length of the series), and 3) both had the same variance (variation about the mean over the length of the series).
Fourth, Turner's assertion that if two series were correlated by 0.5, a one-point move in one would be matched by a 0.5 point move in the other is inaccurate. Even if both series had a mean of zero and variance of one, statistical measures of association are really about probability in the sense that traders talk about expectancy. Thus, it would be more correct to say that on average, given a one-point move in A, the best prediction we can make is a 0.5 move in B. But given that we are discussing probabilities, the actual move could be greater or less than 0.5, or even in the opposite direction. And since the conditions of equal mean and variance are almost never met, the actual predicted value needs to be adjusted for these differences as well. This adjustment in the predicted value is done through the procedure known as regression and not by correlation. Even using regression, the expectancy nature of the predicted value necessitates trading discipline and money management techniques.
Fifth, Turner makes passing reference to movements expressed in percentage rather than dollars. Correlations among data expressed in percentages entail nasty statistical problems because of certain formal properties of the percentage transformation (the mean and variance are related and the range of the variable is restricted). While there are data transformations (the arc sine square root being one) that resolve these problems, readers would do best correlating series expressed in dollars or any other currency and avoiding percentages.
Michael Turner replies:Robert Philip Weber, Ph.D., via E-mail
Menlo Park, CA
I am delighted that you took the time to evaluate my article. You make many accurate points, albeit ones with a bias toward academic mathematical definitions rather than from a perspective of general securities trading.
But the concept of correlated stocks is not meant to be taken literally. Rather, it is a metaphorical "equation" that defines the relationship between security prices. Here are my responses to each of your points:
Regarding your first point on removing the trend by differencing, it's true that this is another way to view the data series. The result may or may not be more accurate.
Regarding your second point on lead and lag in time series relationships, please review my article, specifically page 32 of the July issue, where I addressed this.
Regarding your third point, you are technically correct that if two series are perfectly correlated, then an increase in one would not necessarily be matched by an increase in the other. But it was not my intention to take the explanation too mathematically. Rather, this was a metaphorical example of how stocks move in relationship to one another. On your fourth point, again, this example was meant as a metaphor, not a mathematical law.
Regarding your fifth point on the dangers of expressing data in percentages, you are correct in your mathematical assertions. In fact, my article specifically refers to point (dollar) movements. Please review the article for confirmation on this.
DAYTRADING
Editor,Several of the interviews that have appeared in STOCKS & COMMODITIES and other interviews with high-ranking traders I've read elsewhere claim that daytraders (especially those who only began daytrading in the last two to four years) will fail and lose their money when the bull market ends. This claim comes from the theory that daytraders are too inexperienced for bear markets and only know how to trade bull markets. But I think not!
I think this is bitter thinking from the experienced trader to predict that the average daytrader will become a high-risk loser in a bear market. The benefit of daytrading is to close out your position at the end of the day. Thus, each day brings a new outlook to the market, whether it be bullish or bearish. I think daytraders can as easily and comfortably short the market as they can buy it; if not, they are probably not actually daytrading.
I agree that daytraders are accustomed to bull markets, but what trader or investor isn't at this point in market history? When the market turns, it will be unexpected and everyone will get burned to some degree or another -- not just new daytraders!
Technology has advanced to the point of accelerating the learning curve of trading, which is why trading has become so popular, and of course, a bull market also helps attract would-be traders. With all due respect for older traders, I think this attitude that says "all daytraders are inexperienced" is out of line. If a more experienced trader told me that I was going to lose all my money because I didn't know how to trade, I'd be just that much more willing to prove him or her wrong!
Houston Jayne, via E-mail
When the market turns, we hope that all our readers -- both experienced and new traders -- will be armed with tools to prepare them for any possible change in the market. Technical analysis can help with trade timing, both entry and exits.-- Editor
TRADE MANAGEMENT AND MENTAL PREPARATION
Editor,I thought your interview with Van K. Tharp in the April 1999 issue of STOCKS & COMMODITIES was fantastic and have since read his book, Trade Your Way To Financial Freedom. You should get him to contribute more articles to STOCKS & COMMODITIES. Failing that, could you emphasize money management and risk analysis more?
Van K. Tharp has been contributing articles on trading psychology to STOCKS & COMMODITIES since 1987. Check our S&C on CD Index to locate those articles, or use our Website's search engine at www.traders.com.Wolfgang Rebien, via E-mail
Welshman's Reef Victoria, Australia
Thank you for your feedback and topic suggestions.-- Editor
FUND TRADING SYSTEM
Editor,I loved the article "A System For Trading Fidelity Select Funds" by Jay Kaeppel in your July 1999 issue. Loved it because I use a similar system with great success trading other fund families. I use slightly different indicators, but the principles are the same. It put me in the European funds in late 1997 and took me out just when the financial press was touting them as they topped out. It put me in large caps, then put me in bonds before the selloff of 1998. As the market bottomed in late 1998, I was taking profits in bonds and moving back into stocks for the rally. Lately, I've been in Asian/Pacific Rim funds.
Specifically, I screen the fund family first with MetaStock Explorer using three weighted moving averages. Then I do a visual screen of relative strength lines in a custom template comparing the strongest fund to all other funds in the family. It works well. It's simple because it's based on price and price alone. The simplicity reduces the probability of system failure and being on the wrong side of a major move one way or the other. Without being specific, the rate of return achieved justifies the effort. It exceeds the buy and hold return, and naturally, exceeds returns achieved by timing services trying to catch tops and bottoms.
Joseph Cavallo, via E-mail
ENDPOINT FFT
Editor,I would be interested to know if anyone has successfully coded the endpoint fast Fourier transform system as described in the May 1999 S&C article, "The Endpoint Fast Fourier Transform System."I have results that are vaguely similar to the published trades for the system, but they do not reproduce it exactly. Anyone who would like a copy of my Visual BASIC code is welcome to E-mail me at Rod@rodx.demon.co.uk.
Rod Gibson, via E-mail
INVESTMENT SOFTWARE INDUSTRY
Editor,I am looking for current information pertaining to the investment software industry. I have a copy of your S&C on CD, from which I located two articles, though they are not current. Do you have any current articles regarding this industry? If not, would you be able to suggest where I might be able to find this information?
Since we are a how-to magazine that presents trading techniques, we don't usually cover general industry news except in our Trade News & Products column, which presents updates on products and services related to trading. However, we do publish reviews of software and products every month, and our Website, www.traders.com, offers a search engine covering many industry Websites.-- EditorKirsten Ebine, via E-mail
San Pedro, CA
CAREER ADVICE
Editor,I have been a reader of your magazine for two years now and am very interested in pursuing a career in technical analysis. As I have just graduated from college, I was wondering if you had some ideas about the career path that I should take to get started in this field. I appreciate your thoughts.
If you're interested in obtaining certification in technical analysis, the Market Technicians Association (MTA) administers a program for obtaining certification as a Chartered Market Technician (CMT). Contact the MTA:Jim Moulison, via E-mail
MTAIf you wish to pursue formal education in finance, trading and investing, The New York Institute of Finance (NYIF) offers applied training and education in a variety of formats, such as classroom training, seminars and conferences, customized training, independent study and exam preparation. Contact the NYIF:
World Trade Center, Suite 4447
New York, NY 10048
E-mail: ShelleyMTA@aol.com
Internet: www.mta-usa.org
New York Institute of FinanceIn addition, several universities in New York City also offer an assortment of finance courses and programs, including Pace University and New York University, among others. Contact any of these schools for a course listing.
2 Broadway, 5th Floor
New York, NY 10004
Phone: 212 859-5000 or 800 227-NYIF (6943)
Fax: 212 344-3469 or 212 514-8423
Internet: https://www.nyif.com
Out on the West Coast, the Institute for Technical Market Analysis of Golden Gate University in San Francisco offers a curriculum of graduate-level classes that lead to a master's degree in finance. Those classes cover technical analysis topics. Contact:
Dr. Henry PrudenTo pursue a career with a stock brokerage, apply to a stock brokerage firm registered with an exchange to enter a stockbroker training program. To become a commodity trading advisor (CTA), contact the Commodity Futures Trading Commission (CFTC) for information, or try reading Trade Up by Holliston Hill Hurd, which provides advice on becoming a professional commodity trader or advisor. To purchase a copy, contact:
Institute for Technical Market Analysis
Golden Gate University
phone 415 442-6583
fax 415 442-6579
Internet: https://teleport.com/~ifta/TSAA/tsaahome.html
Futures TruthIf you are interested in preparing articles for STOCKS & COMMODITIES on technical analysis, send for our Author Guidelines or visit our Website. Our guidelines cover topics and remuneration.-- Editor
815 Hillside Rd.
Hendersonville, NC 28739
828 697-0273
E-mail: FuturesTruth@a-o.com
Internet: https://www.futurestruth.com
AUTHOR THANKS
Editor,Thanks for doing a great job of publishing my "Yield Spread Tunnel" article in the June 1999 STOCKS & COMMODITIES. I really like how it came out -- the color, the artwork, the graphics, everything. Please extend my thanks to everyone involved.
Lorne Rae
Vancouver, BC, Canada
ERRATA
In our July 1999 Books For Traders section, we listed an incorrect fax number for book publisher M. Gordon. The correct fax number should be 213 955-4242.Back to September 1999 Contents