November 2002 Letters To The Editor
or return to November 2002 Contents
The editors of S&C invite readers to submit their opinions and information on subjects relating to technical analysis and this magazine. This column is our means of communication with our readers. Is there something you would like to know more (or less) about? Tell us about it. Without a source of new ideas and subjects coming from our readers, this magazine would not exist.
Address your correspondence to: Editor, STOCKS & COMMODITIES, 4757 California Ave. SW, Seattle, WA 98116-4499, or E-mail to firstname.lastname@example.org. All letters become the property of Technical Analysis, Inc. Letter-writers must include their full name and address for verification. Letters may be edited for length or clarity. The opinions expressed in this column do not necessarily represent those of the magazine. -Editor
WAITING FOR THE FED
Can you clarify a point in the article "Waiting For The Fed" by David Penn, which was published in August 2002 issue of STOCKS & COMMODITIES? Does Mark Boucher's system keep out of stocks if the T-bill rate is lower than the prior year? I got confused in reading the description of the system and the commentary.
Glenn Bergevin, via e-mail
David Penn replies:
Thanks for writing. The monetary timing model I mentioned is a rate of change model. In other words, it doesn't matter if the percentage change is up or down. If the year-to-year change is greater than 6%, suggests the model, then an exit signal is given. If the year-to-year change is less than 6%, then an entry signal is given.
I think this is key. Boucher points out the conventional wisdom that higher short rates are bad for stocks. In my article, I wanted to suggest that the problem isn't just higher or lower short rates, but the rate of change from year to year. Stocks, apparently, can thrive in both relatively higher and relatively lower short rates. But what appears to really affect stocks badly is when short rates change dramatically from year to year.
I hope that helps. We should see some interesting evidence if and when the Fed begins to raise rates (or until enough time passes and we have two years of historically low interest rates in back-to-back years).
Thanks again for writing, and for reading Technical Analysis of STOCKS & COMMODITIES and Working Money magazines.
I am amazed at how Spyros Raftopoulos misunderstands zigzag in his August 2002 STOCKS & COMMODITIES article, "Zigzag Validity." On page 28 of that issue he states, "In fact, you could not even tell if the closing price that day was really a trough; you'd have to at least wait until the next day."
The correct interpolation of zigzag is: If a 10% filter (threshold level) is used, a trough can only be confirmed after an upswing that exceeds 10%, no matter how long it takes. If, for example, the next upleg takes two months and has a maximum upturn of 9.9%, it is not confirmation, as it could well turn around and drop 50%, creating a potential new trough and negating the old one. It therefore takes a complete upswing exceeding 10% to lock in the previous drop (trough).
His followup statement, "This is an unavoidable weakness of the zigzag indicator; since it depends on peaks and troughs, it should always identify every new swing point with at least a one-bar delay, which is the time needed for reversal confirmation." Wrong again for the same reasoning as above: It takes a followup confirmed cycle to lock in the previous cycle under consideration, no matter how long it takes.
On page 29, he is more inaccurate when he states, "Further, when you look at zigzag on the bar chart, you get the impression that the indicator has accurately predicted every major move of the past and has avoided the minor ones, which is not true at all." First, zigzag does not "predict" anything; it determines up and down swings of a given amount well after the fact and it does so with complete fidelity (within the ground rules of zigzag). Second, it has not avoided the minor moves; it "sums" them until they accumulate to a level that determines a peak or a trough.
In Figure 2, "peak C" cannot even be considered a peak (as it has less than a 10% upswing) any more than the lesser peak immediately before it. One simply has to wait for the up excursion to exceed 10% before conditionally classifying it as a peak, as only a subsequent downswing of more than 10% can validate it as a true peak. Thus, zigzag is always "running" one cycle behind, which doesn't negate the information it is telling us; it simply must be interpreted correctly.
Norman J. Brown, Danvers, MA
Spyros Raftopoulos replies:
You have a good perception of the zigzag function. The problem is that you fail to see mine. Your disagreements are due to misreading my article; most of the comments you write are similar to the arguments I make in my article. In addition, the code for my zigzag validity indicator is included in the sidebar to my article on page 30. If you had tried this indicator, you would see that it correctly validates the last leg of zigzag.
As for the two statements of mine that you quote: "In fact, you could not even tell if the closing price that day was really a trough; you'd have to at least wait until the next day," and, "It should always identify every new swing point with at least a one-bar delay," essentially mean the same thing - that is, you cannot see a trough if the prices don't turn up first. In my example, I used a daily chart of Ibm and the zigzag of closing prices. In such a chart, you aren't able to observe any reversal (whether an important one or not) in less than one day. The above statements of mine do not imply that reversals of any kind are time-dependent, and nowhere in my article do I make such a statement.
Further, when you say, "The correct interpolation of zigzag is: If a 10% filter (threshold level) is used, a trough can only be confirmed after a following upswing that exceeds 10%...." Here, you are repeating in your own words what I have expressed in various ways in my article. Take for example this sentence of mine: "The reason the B-C move was considered unimportant was that the change from B to C was less than 10% (this sensitivity threshold is set by the user)." It is obvious that we both are describing the same thing, the difference being that your expression is slightly inaccurate: a trough is not confirmed after a following upswing that exceeds 10%. It is confirmed on the very day that the closing price exceeds the 10% threshold. But this often happens long before the formation of a new upswing!
I agree with your remark that zigzag does not predict anything. I am simply talking about the impression one may get by looking at the zigzags. My statement is very clear: "You get the impression that the indicator has accurately predicted every major move of the past and has avoided the minor ones, which is not true at all."
I also agree that zigzag has not avoided the unimportant moves. Again, I am talking about the impression one may get by looking at zigzag formations. Actually, the dynamic nature of the last leg of zigzag provides clear evidence that the indicator cannot magically avoid any unimportant moves. In order to filter out the minor moves, it first has to deal with every single one.
This also responds to your comment: "In Figure 2, peak C cannot even be considered a peak (as it has less than a 10% upswing) any more than the lesser peak immediately before it." So, one has to decide: Are they peaks or not? My answer is clear: they are peaks, beyond any doubt! Swing points were formed long before the invention of the zigzag indicator. In fact, every time prices change direction, they form peaks and troughs. What is indicated by zigzag is their importance, only.
Unfortunately, the last leg is often misinterpreted. I hope that my zigzag validity indicator, if used correctly, will help prevent these misinterpretations.
I've worked out an active trading strategy that seems to perform well. I have four years of experience in actively trading Nasdaq stocks. I'd like to share my trading approach and experience and I am looking for funds to manage. Do you organize any kind of contest between active daytraders? If not, maybe you could give me some advice what I should do in my case.
Roman Kulyan, via e-mail
We do not organize or host any trading competitions, but here are a few trading contests to mention:
Several other competitions or contests are listed in this article:
You can locate additional trading competitions by searching the Internet. We are not affiliated with any of these competitions, nor can we make any recommendations about them. Some of these services may have been reviewed by STOCKS & COMMODITIES. Past reviews can be located through the search engine at our website, Traders.com. -Editor
I really enjoy your magazine each month. It has been very helpful in the development of my trading systems. Keep up the great work.
I have some confusion regarding the measurement and definition of volatility. I'll give two extreme examples (shown in Figures 1 and 2). It seems to me that a market can be very volatile while still having very nonvolatile prices (a low average true range) (Figure 1). On the other hand, a market can be "flat" while having large variations in daily highs and lows (volatile average true range) (Figure 2). According to the average true range indicator, Figure 2 represents a more volatile market than Figure 1. At first glance, however, most readers would consider the chart in Figure 1 as the more volatile. Price volatility and market volatility appear to be two different things. What indicator can be used to gauge market volatility?
Gary Mellor,Calgary, Canada
Theoretically, the sample charts you have provided may be correctly interpreted, but I found it difficult to recreate a similar scenario using actual price charts.
The average true range (ATR) is an appropriate indicator to use for analyzing volatility. I looked at various charts and found the ATR hitting lower levels during flat markets. However, if you have concerns, there are various other indicators you can use to measure volatility, such as Chaikin's volatility indicator, Bollinger Bands, and standard deviations. Another option is the volatility index (VIX), which measures volatility in the options markets. You may wish to look at articles we have published in the past on these topics. You can locate them by using the search engine at our website, Traders.com.
Hope this helps.-Editor
I have been on the learning curve with commodity trading for about 16 years now. I came up with a trading system about 14 years ago that has done very well. (At that time, I was trading the S&P and bonds.)
It's now about 30,000 hours later and I've continued to develop that original system, so that it is now four separate systems that can be traded together in order to smooth out the curve.
When the S&P margin got too high, I left it and went looking for another market to trade, and ended up with the Swiss franc (which is an inverse proxy for the US dollar index, taking its tops and bottoms literally on the same day as the dollar index, except in reverse).
Finally, when overnight trading sessions were introduced, I repeated the same two original systems for the overnight market, folding the four together.
Never content, I kept going and evolved two different moving average crossover systems (built a little differently from the conventional), and these can be folded into the system to make six systems that are working together at one time.
I think the reason most systems only work for a while is that they are developed during a trend and work fine until the trend changes or the market flattens out (which occurs possibly 85% of the time). What I discovered is that the stops that I interpolate in a nearest-contract situation will not work when I have to rebuild my template for the next nearest-contract situation when it becomes the contract most traders will be in.
That, I think, is the key: Rather than insisting that the market move to me, I let the market do whatever it is going to do, and adapt to its movements (which may be repetitive in general, but really it's never quite the same in a day-by-day environment).
Alan Scott,Issaquah, WA
Thanks for your comments about trading system development. If you ever feel inclined to share your methods with other traders, we would be happy to work with you on an article. Good luck with your endeavors.-Editor
ERRATA: METASTOCK CODE in "DEVELOPING A TRADING SYSTEM"
I have been a subscriber for several years. Never have I seen such a mess as the code given for MetaStock in the sidebar on pages 50-51 of the August 2002 S&C in the article "Developing A Trading System."
You should start again and publish accurate code, being sure to tell the reader where to input it. Simply transferring the script from the original indicator position to the Explorer does not work. What is needed are the details of the boxes each part needs to be copied to.
George E., via e-mail
Regarding the MetaStock code given in that article, we had several readers write in to let us know that the furnished MetaStock code was not usable.
Because of MetaStock's limitations, the code given in the article was not the actual MetaStock script used. But it did demonstrate what author Dennis Peterson wanted to do, and it was something that he was able to program in Wealth-Lab instead. That Wealth-Lab script was given in the same sidebar.-Editor
Back to November 2002 Contents