or return to February 2001 Contents
PHASOR DISPLAYS
Editor,
I don't like to be negative, but I must admit that John Ehlers's article in the December 2000 S&C, "Phasor Displays," as usual, frustrates me. On a scale of one to 10, I'd grade him, brilliance: 9, clarity of exposition: 1.
For example, a small puzzlement: In Figure 7, in quadrant II there's a note, "point 28." Isn't this point 23? There's a point 28 near the top of the chart, in quadrant I. In quadrant I, there's a note, "counter clockwise segment." Isn't the swing bar 21, 22, 23...41 a clockwise swing?
It could have been helpful to include a chart showing typical trend modes and another showing typical cycle modes.
Keep up the good work -- the magazine is great.
Art Merrill
Haverford, PAGood to hear from you, Art, as always. Don't be too hard on Ehlers; I mash up his input pretty good before it gets to you. You are correct about point 28 being point 23 (Figure 7); my mistake. However, you are wrong about the counterclockwise notation. There is an arrow pointing to the short segment that is going counterclockwise instead of clockwise. And thanks for the idea about having a chart showing typical trend modes and another with typical cycle modes; I'll ask Ehlers to do something along this line. Isn't the display spectacular!? -- Editor
MOON CHART MADNESSEditor,
Several letters to the editor have touched on whether there is any link between the market and the moon. If any connection exists, it has little to do with astrology as it is popularly conceived. Either there is an effect or not, depending on whether there are physical or psychological forces at work. However, popular mythologies may sometimes contain anecdotal clues to puzzling aspects of nature. Astrology was one of the precursors of astronomy, just as alchemy was a forerunner of chemistry.
The question of whether there is a lunar influence on the stock market could be rephrased as, for example: Does the increasing gravitational pull of the moon aligning with the sun have an effect on the Earth? Clearly, it does (in the form of tides). Might this effect also extend into subtly affecting human behavior, such as the sometimes emotionally laden decision to buy or sell in the stock market? Posed in such a manner, this could be an interesting and reasonable question to ask.
I was somewhat intrigued by the chart shown in the December 2000 Letters To S&C (Figure 1: Full Moon Vs. Rising Days In The Dow Jones Industrial Average). The caption of that interesting graph from Arthur Merrill implies there is no significant correlation. However, if you look at the trend of the chart's data over the four days immediately preceding the full moon (days labeled 1 through 4), there is a consistent rising pattern that may in fact be significant and correlate well with an increasing gravitational/tidal pull. An auto-correlation measuring the percentage change from one day to the next for that time segment might point to an increase in buying decisions or at least to higher price volatility.
I have been involved in computerized analysis of stock market data for nearly 10 years. And although I am quite hard-nosed and skeptical of many technical analysis claims, I have learned that there do seem to be some interesting effects that are not easily uncovered using standard statistical measures. Such effects, if any, may be subtle and/or difficult to measure. My main point concerning whether there is a "market in the moon" effect is that the issue is worthy of continued exploration.
Richard McKee
Washington, DCGary suggests you try a different broker. -- Editor
TRADING SIMULATORSEditor,
I wonder if you know any futures and commodity trading simulators using real-time data. I'd like to test-trade before jumping in.
Robert Tyminski
Brooklyn, NYWe're not aware of any. Real-time datafeeds require exchange fees, which you would need to pay. -- Editor
BROKERAGESEditor,
Do you know of a brokerage through which I can trade both stocks and commodities from the same account?
Lyle Ziegelmiller
via e-mailAny full-service broker -- Merrill Lynch, for example. -- Editor
BUY/SELL RECOMMENDATIONSEditor,
As a new subscriber, I have checked out your website, graphs, quotes, and charts, but I have been unable to find buy or sell recommendations. How can I access this information?
Bob Wilson
via e-mailIn order to give such recommendations, we would have to be a registered commodity trading advisor, which we aren't. Thus, you won't find any buy or sell recommendations on our site. -- Editor
TAX SEASONEditor,
I have enjoyed reading STOCKS & COMMDOTIES for the last six months. I find it to be very interesting and useful. As tax season approaches, I would like to see some articles or book reviews on the subject. I think understanding the taxation issue can be valuable to many traders.
Chaim Pazerkovsky
New York, NYThank you for the suggestion. In the meantime, see Ted Tesser's article, "It's What You Keep That Counts," in the January 2000 issue of S&C. -- Editor
TROUBLE WITH RSIEditor,
I am an Italian trader on LME and COMEX exchange. I read your magazine with pleasure, and have found a lot of good tools. If possible, I need a little help.
In trading commodities (in particular, copper, ALI, zinc, and sometimes nickel), I use MetaStock 7.0 Professional. I receive data via satellite. I would like to create an indicator that links a 14-day RSI and the variation of daily close prices. For example, if RSI is at 24 on a 60-minute chart of copper: I have to buy! But in periods where the market doesn't find a direction, RSI is not a valid indicator because it gives indications that are not in line with the real variation of the quotes. I need an indicator of sensitivity. If you have some ideas, please tell me.
Enrico Grossi
via e-mailYou've run across a big problem with oscillators. Look into VIDYA (variable index dynamic average), developed by Tushar Chande ("Adapting Moving Averages To Market Volatility," STOCKS & COMMODITIES, Volume 10: March). -- Editor
S&C ON CDEditor,
I want to be clear: Does your $395 S&C on CD include every single article and all article images (charts, tables, and so on) for all the STOCKS & COMMDOTIES issues since Volume 1 and through 1999? I have purchased a subscription to your magazine, but I want all your past issues as well. I need to know exactly what issues the CD contains (and that it contains everything from the issues included), as well as what issues it doesn't contain. I would also like to know when the 2000 issues will be included on the next CD-ROM and what the cost is to upgrade. Also, do you offer updates to the CD-ROM annually? Thank you in advance for your time and help.
Erik Weaver
via e-mailOur S&C on CD contains all articles from S&C since 1982, including art, charts, tables, and code. Some regular features of the magazine were purposely left off of the CD, such as the Editor's Opening Position, Futures Liquidity, and Trade News & Products, since these features are somewhat timely in nature and do not have much use in the long term. The CD also includes some Excel spreadsheet files from some articles. No ads were included, only editorial material. The current CD includes articles from 1982 through December 1999. The next version will likely come out by early February and include articles through December 2000. Upgrades are $39.95. -- Editor
SINGLE-ARTICLE COPIESEditor,
I would like to know how to get a copy of the article "Andrews Line" from the October 1996 S&C by Ron Jaenisch and the article "Support And Resistance With The Andrews Pitchfork" from the November 1995 S&C by Barbara Star.
D. LaRocco
via e-mailCopies of individual articles may be purchased from our Online Store at our website, www.traders.com. -- Editor
THE BAD-FORM BEAREditor,
After working as an associate for an investment bank performing fundamental equity research for two years, I discovered and became interested in technical analysis. I discussed my newfound interest with a friend, and he gave me a couple of old copies of STOCKS & COMMODITIES to look over in my free time. As I sat down to read them, before I even opened the magazine, I was struck by an error.
On the cover of your August 2000 issue, a stock market bear, whose name on the uniform I'm unable to make out because it is partially out of view, is pitching in the middle of a sold-out All-Star baseball game. I say "sold-out" because there does not seem to be an empty seat in the house, and I say All-Star because, first, it is a summer issue, which probably came out around the time of the 2000 All-Star game in Atlanta this past season, and second, it appears as though two bulls playing in the field behind the bear are wearing different colored hats.
I played baseball for most of my childhood and throughout college, so I can tell the bear has good form. He has good balance, his eyes are on the target, and he is hiding the ball very well. The only problem I see is that he, being a right-handed throwing bear, has the wrong knee in the air as he attempts to pitch.
This situation would make it very difficult for the bear to get any power behind the ball as he throws his pitch, and would not make for a long career considering the arm problems he would sustain from doing this for an extended amount of time. How the bear made it to the All-Star game with such an error in his pitching motion is beyond comprehension.
Despite this fact, I do enjoy your magazine and look forward to future baseball pictures and sketches your art staff can come up with. This only brings up one other thought: Had the bear been a left-handed throwing bear, this would have given new meaning to the term southpaw.
Keith C. Applegate
New York, NY
MARKET TOPOLOGYEditor,
I really appreciated the article by Nicolas Vandewalle in the December 2000 S&C, "Market Topology." I believe you would boost the circulation of your magazine if the cross-correlation coefficient tree were published in every issue for the previous month, maybe in a two-page chart. I find it very helpful.
I have a few questions. What was the length of the time series? 13 days? 26 days? A more important question: What is an appropriate time series length? I imagine there comes a point where the tree becomes unstable if the time series is the wrong size. What size does the job best?
Also, how is a cutoff determined? Are the coefficients sorted with anything more than a statistical measure? What makes a good cutoff?
Great magazine. Keep up the good work.
Reginald Creighton
via e-mailNicolas Vandewalle replies:
Thank you for your interest in our work at www.market-topology.com. The time series we analyze are one-year long, that is, around 250 days. For such series, the cross-correlation coefficients we measure are quite relevant. It is also possible to reduce the time series length to six months without drastically changing the results. However, less data points would lead to irrelevant random trees.
Concerning the cutoff in the distribution of cross-correlation coefficients, for N stocks, the number of all possible pairs is N(N-1)/2. Some pairs are uncorrelated (negative correlation), and others are highly correlated (positive correlation). There is indeed a broad distribution of coefficients.
From this distribution of N(N-1)/2 coefficients, we select the N-1 most relevant ones for building a tree-like topology. The tree is assumed to represent the skeleton of the market. Two working conditions are thus imposed: building a tree (loops are avoided), and selecting relevant coefficients. With those conditions, a natural cutoff appears in the distribution of coefficients selected for the tree. The cutoff depends mainly on the correlations lying in the market. For example, the cutoff for the US market is around 0.2, since the tree shows pairs of stocks with correlation coefficients larger than 0.2.
Editor: Unfortunately, we haven't the space to publish these intriguing diagrams regularly. I would encourage Vandewalle to publish them on www.market-topology.com or the University of Liege website!