October 2002 Letters To The Editor
or return to October 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 editor@traders.com. 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
FAST FOURIER TRANSFORM
Editor,
The recent article "Fast Fourier Transform" by Amy Wu in the July 2002 STOCKS & COMMODITIES was an informative article. The last two sentences stated: "...many aspects, such as phase, are being discounted to fit the type of data in use. [I agree.] Still, interpreted correctly, FFT can greatly enhance your accuracy on other more fitted indicators."
We need a followup article telling us what these other indicators are - such as stochastics, MACD, RSI and others - and how to select the best parameters for them using the FFT results.
Lee S. Akin, via e-mailThank you for your topic suggestion. We have passed it along to Amy Wu.
Readers wanting to read more on fast Fourier transforms and Fourier analysis can visit the Online Store at our website, Traders.com, for past STOCKS & COMMODITIES articles. In fact, this topic was one of the very first studied in this magazine when it was launched in 1982 by our publisher, Jack K. Hutson.-Editor
ONLINE INVESTING DISCUSSION GROUP
Editor,
I am interested in starting a discussion group on online investing. How would I go about doing so?
William M. Frantz, via e-mailYou can visit our Message-Boards at https://Message-Boards.Traders.com to discuss trading ideas with other readers, and also try searching the Internet for other online discussion forums on investing.-Editor
DEVELOPING A TRADING SYSTEM
Editor,
I have enjoyed reading S&C for several years now. I tried to test the MetaStock code included with Dennis Peterson's article, "Developing a Trading System" in the August 2002 S&C.
However, I have MetaStock 7.2, which will not allow me to insert a variable into a LLV, HHV, or BBandBot function. Does he happen to have a cunning approach he would like to pass on, or have I another problem?
Charles Schwedes, via e-mailStaff Writer Dennis Peterson replies:
Unfortunately, I know of no cunning approach, and it is the reason I ended up going to Wealth-Lab. If you find a way, please let me know because there are others who would like to know as well.
METASTOCK CODE FOR SYSTEM DEVELOPMENT
Editor,
I enjoyed reading Dennis Peterson's article "Developing A Trading System" in the August 2002 STOCKS & COMMODITIES. He made some very interesting points, and I was able to glean several good ideas from the article.
But I experienced several problems trying to implement the system test to validate his results. In my opinion, the sidebar for the MetaStock code was written in a confusing way and did not allow easy implementation. Here were the problems I encountered:
1. Peterson mentions in the sidebar that he would like to be able to compare the volume on a signal day to the volume on the last up day. This is an easy formula. In MetaStock code, the condition could be written as follows:
V>ValueWhen(2,C>ref(C,-1),V)2. BBandBot does not allow a "variable" such as BBpds in the construction of its formula; thus, you receive an error message when you attempt to use the formula as written in the sidebar. I had to substitute a constant value (I used a value of 16 as an average value) to allow the system test to compile.
3. In the "Exit Conditions," there is a custom formula labeled "StochRSIvol," which is defined but never used.
4. There are more lines of code in the "Entry & Exit Conditions" than is allowed in the entry and exit sections of MetaStock's System Tester. I had to substitute formula references for the StochRSI to get the code to fit in the allowed space.
5. The formula for the StochRSI was buried in the exit conditions twice, taking up additional lines of code.
After several hours of work on the formulas in the sidebar, I was able to get the system test to compile, and I came close to Peterson's results. My concern, though, is that the system test is suspect since I had to make so many changes to get the test to work. - Robert Frassanito, via e-mail
Staff Writer Dennis Peterson replies:
1. Thanks; I didn't know about the "ValueWhen" function, but it does appear to do the job.
2. You have to set BBpds to be equal to one of the optimization variables, such as opt1.
3. The StochRSIvol variable was used in my first attempt to incorporate volume. I can't say that it won't work in a modified form - I would think it would - but I ran out of time before I could get it to work, so I ended up leaving it in for reference. I did use it at one point, but I didn't like the results and wanted to come back to it.
4. As for the lines of code allowed in the MetaStock System Tester, I did run into this problem as well, but the code in the sidebar is a direct capture from MetaStock, so I am a bit puzzled as to what is going on.
5. As for the formula for the StochRSI that was given twice in the exit conditions, yes, I see that and apologize for having it in there twice - it obviously only needs to be in there once.
As for getting the system test to work, please keep in mind that almost all the system tests were done by optimizing on periods and thresholds. The code shown was the final attempt to get to something that tried to be variable in nature.
TRADING IBM INTRADAY
Editor,
As a relatively new subscriber, I've found your magazine to be invaluable in my education about daytrading stocks and about the markets in general. Relating to the article "Trading IBM Intraday" by Dennis Meyers in the June 2002 S&C, however, I have several comments.
First, this article is presented in an esoteric manner and never gives details as to how the "system" works. It gives references to other places where the reader can find more information - but these places are not easy to access, particularly sitting in front of your computer, reading S&C, and wanting to test strategies presented.
Second, the sidebar "Mathematics" presents theory, but no details, as to how this type of system is implemented.
Third, in the Traders' Tips section of that issue, Marge Sherald of Ward Systems addresses Meyers' article. First, she mentions that the article provides insufficient information for the reader to implement (as noted above in my comments). Then, she goes on to present an unrelated neural net system that gives completely different results (better, of course!), and only vaguely refers back to the article for comparisons and benchmarks.
Fourth, again in Traders' Tips, page 79, Dion Kurczek presents results for Wealth-Lab using an EMA ("For simplicity's sake, we'll use the first-order polynomial (an EMA) instead of the fourth order"), again providing a Traders' Tip that does not implement anything near to what the article was intending to present.
This article may be more appropriate for a mathematics magazine, since it does not present sufficient usable information for anyone to actually implement the system. Instead, the article presents a system that sounds like the holy grail for IBM (and by implication, applicable to "other markets or stocks"), and doesn't provide any means for the reader to test, modify, or apply the system. The system is not verifiable by the reader; all results presented must be accepted on faith. As such, I think it was a waste of space in your magazine. Your articles are generally targeted at the real-world trader (as they should be), not at theoreticians.
Further, your Traders' Tips section should provide usable code and algorithms. I'm a TradeStation user, so EasyLanguage code would be nice, but code for any platform (MetaStock, eSignal, NeoTicker, etc.) provides the algorithms and calculations necessary, so that I can easily convert it to the format I need. Both of the Traders' Tips provided for this article, written by experts for their respective platforms, were unable to implement the system presented in the original article - and certainly, if these experts could not do it, then your typical reader will have no chance.
Thanks again for the great info you present each month. I read cover to cover. - Nelson Crowle, via e-mail New Port Richey, FL
Thank you for your feedback. We try to present a range of articles in each issue of STOCKS & COMMODITIES, some for every level and type of trader. Not all articles will be useful and applicable to every reader's trading.-Editor
WAITING FOR THE FED
Editor,
Thanks for David Penn's Working Money article published in the August 2002 issue of STOCKS & COMMODITIES, "Waiting For The Fed." I'm always searching for profitable timing models. I think I may need a little help with this one.
In the article, Penn discusses hedge fund manager Mark Boucher's monetary timing models. I was wondering, does Boucher use constant maturity or secondary market rates from 1947 to 1997? Are negative integers treated as positives when calculating ROC? Average annual gains may, at times, be a bit misleading. Were there many losing trades and were any of them rather large, say 15% or more? - Les Brod, via e-mail
Staff Writer David Penn replies:
Boucher doesn't specify in his book The Hedge Fund Edge which type of three-month Treasury bills he's using - which leads me to believe that it doesn't matter a great deal as long as you are consistent. In my "Waiting For The Fed" article, I used constant maturities as found on this section of the Fed's website:
https://www.federalreserve.gov/releases/h15/data.htmAll values are absolute values with regard to rate of change (ROC). This is the most important part of the model, it seems to me, and I'm not even sure that Boucher was fully aware of it. The conventional wisdom is that when the Fed starts cutting rates, the smart money starts buying stocks. I argue that what is more important is that short rates stay relatively calm (a variance of less than 6%) from year to year. Stocks can appreciate in environments with short rates much higher than we would ordinarily be comfortable with. What gives stocks difficulty is high volatility in short rates. Someone who started buying stocks when the Fed started lowering rates in 2000 would have been stuck with miserable returns (the Dow Jones Industrial Average lost 5.6% over that period; see my article "The Easing Season" beginning on page 100 of the July 2002 S&C for the discussion). Boucher's model would have kept investors or traders out of the market for most of 2001 and all of 2002, to date.
Drawdowns are an important point, I agree. This is another area where the model impresses me. In Boucher's survey from March 1947 to October 1995, the worst period was a 14.59% loss from March 1970 to May 1970. There were only seven losing trades in that 48-year period, with the average loss (including the worst above) being only 6.18%. My article shows that since 1995, the only losing trade was very small (a 2% loss in the January?February 2001 trade).
Interestingly, after that losing trade from March 1970 to May 1970, which lost 14.59%, the next trade was a buy in July 1970 with a sell in March 1972 for a 47.41% gain.
Boucher would also agree with you that annual rate of return can be misleading, and he encourages using compound returns throughout The Hedge Fund Edge (which is an excellent book, by the way). He observes:
"The reason we use compound annual return instead of average annual return is that average returns fail to show as vividly the effects of large negative years on long-term returns when compared with compound annual returns ..."
Thanks for writing.
STILL Waiting For The Fed
Editor,
I was very interested in your Working Money article "Waiting For The Fed" published in the August 2002 S&C. I have been following Marty Zweig's models for a few years now and have been a bit disappointed in its recent track record (flashing a buy from February 2001 through to today's [July 2002] miserable lows). The Boucher model, by comparison, looked very interesting, especially when you mentioned that "the model gave a buy signal at the same time the Fed began lowering the Fed funds rate in January 2001, but got back out of the market a month later with a 2% loss." I have tried to recreate the Boucher model on my own but have come up short. I was wondering if you could review my results?
You wrote that "the model calls for buying the Standard & Poor's 500 when the year-to-year rate of change in the yield of the three-month Treasury bill is less than or equal to 6%." I looked up the yields on the three-month T-bill (FRB report H.19, "3 Month Treasury Bill: Secondary Market Rate"). As an example:
In January of 2000 the rate was 5.32%
In February of 2000 the rate was 5.55%In January of 2001 the rate was 5.15%
In February of 2001 the rate was 4.88%
To find the year-over-year rate of change for January of 2001, I took 5.15% - 5.32% = (0.17)% (therefore, the yield is down 0.17% yr/yr). I then divided by 5.32% to = (3.2%) (therefore, one-year rate of change on the three-month T-bill was a decline of 3.2%); -3.2% is less than or equal to 6%, so the model signals a buy. For February 2001, I did the same: 4.88% ? 5.55% = (0.67%); (0.67%) / 5.55% = (12.07%); -12.07% is less than or equal to 6%, so the model still signals a buy. This goes against your article, which said that the model should signal a sell. I have backtested my formula and the dates seem to correspond with Boucher's dates of change on his model (that is, buying in November 1984, selling in August 1987).You later show an example where "the three-month T-bill yield in summer 2001 hovered between 3.7% and 3.4%. A buy signal would be initiated if, during the summer of 2002, the three-month T-bill yielded something in the neighborhood of 3.9% on the high end, with 3.2% representing a minimal (less than 6%) year-to-year change." From this, it appears that you look to see that the yield one year later is +/- 106% of the previous year. Using that formula, I see that you would exit the market in February of 2001, as you mentioned. However, when I backtest that formula, I don't seem to have any long-term buys other than when rates remain flat for a long period of time.
You mention that Boucher's model would trigger a buy if rates were to rise. Wouldn't rates then have to drop back to their current low levels one year from now to maintain the buy signal? It seems as though the way you describe the model, a buy is triggered any time rates mirror the rates 12 months prior, no matter whether they have risen or declined. Is this right?
As you can see, I am quite confused. Any help you can offer would be most appreciated. - Tristran Eddy, via e-mail
Staff Writer David Penn replies:
One error in your logic I think I may have spotted is where you say, "For February of 2001, I did the same: 4.88% - 5.55% = (0.67%); (0.67%) / 5.55% = (12.07%). -12.07% is less than or equal to 6% so the model still signals a buy." You have calculated a sizable rate of change (12%) from the decrease in the yield of the three-month T-bill from February 2000 to February 2001. That 12% rate of change (Roc) would be greater than the 6% Boucher's model calls for, not less. Remember, the model calls for the absolute rate of change, making no distinction whether or not the change is an increase or decrease. This is, I think, a significant difference.
What the model is really suggesting is that when the Fed makes abrupt changes in the value of money (raising or lowering short-term interest rates dramatically in a year-to-year context), the stock market as represented by the S&P 500 has historically fared poorly.
This also shows the paucity of the argument that falling interest rates are an unqualified good for stocks. This may be especially so in a deflationary environment, as I alluded in another article, "Gibson's Paradox," in the June 2002 S&C.
In regard to your question about the three-month T-bill yield and the buy signal that would be initiated, I am really working backward within the Boucher model in an attempt to answer the question: Where would the three-month T-bill's yield have to be in order for the model to present a buy signal in 2002? With the rate in the summer of 2001 being between 3.7 and 3.4%, summer 2002 rates would need to be within 6% of those yields, above or below. Thus, summer 2002 rates in the neighborhood of 3.9 on the high end or 3.2% on the low end would appear to be necessary for Boucher's model to issue a buy signal.
The 106% rate of change you mention, by the way, would mean more than doubling - or halving - the previous year's rate. That would mean a high end of 8% and a low end of -0.2%!
I am particularly impressed that Boucher's model has kept investors out of the S&P 500 for most of the year - and with good reason, as you've mentioned. With the Fed apparently reluctant to raise interest rates, it will be interesting to see whether or not Boucher's model saves investors more money, or whether the model causes them to miss out on the "big rally" that everyone seems to be waiting for.
Again, if the notion of buying on rising interest rates seems bizarre, take a look at both my "Gibson's Paradox" article mentioned above, as well as Boucher's own comments about deflation and Japan in his book The Hedge Fund Edge. Know also that Pimco's bond guru, Bill Gross, and his gang have been trying to nudge higher interest rates out of Greenspan all year.
Thanks for writing.
OPTIONS SOFTWARE AND PAPER TRADING
Editor,
I will soon be in a position where I can focus on options in my daytrading. I love numbers and this is something that I have always wanted to do. However, I have no experience or training. I am currently getting some basic training in options theory, and it has become very clear to me that I am going to need some type of options software to assist me.
Here is where the problem lies: There appear to be tons of software on the market, but I am too green to be able to evaluate the different systems. Can you point me in the right direction? Also, is there anywhere I can sign up to do paper trades? I work very hard for my money, and I want to do hundreds of paper trades before I risk capital, but I have been unable to find a paper trading service. - Bruce Lyon, via e-mail
See our review of OptionSites.com on page 117. See also the next letter and response.
SOFTWARE SEARCH
Editor,
I am new to investing and have a minimal account. What I have been trying to find is a very good piece of software that will help me search, track, chart, and read the different markets.
I understand how to read charts fairly well and use a few of the indicators. I understand the basics and am willing to branch out and grow. I am learning about options and swing trading right now.
I have seen TCNet 2000, Wizetrade, InvestorToolbox, e*Trade, TradeStation, and a few others. I am confused on what would be a good starting point and what might be good to stick with as I grow.
Do you recommend any software or is there anyplace I can go to get an honest review of the top 15 to 20 software programs? I would appreciate any help you or your team can point in my direction. - D'Arcy A. White, via e-mail
We publish reviews in every issue of STOCKS & COMMODITIES. Our reviews take a detailed look at how the program works and shows the reader what it is like to use the product. You can check when we last reviewed a particular product by using the search engine at our website, Traders.com, and you can buy reviews from our Online Store.
You can also visit the Traders' Resource listing at Traders.com and browse the category of software or trading systems or sort by particular features you desire. Follow up by contacting the individual software vendor for more information.
Finally, you can exchange comments and questions with other STOCKS & COMMODITIES readers at the Message-Boards at our website to discuss product usage with them and get their insight.-Editor
DEVELOPING A TRADING PLAN
Editor,
I have enjoyed reading S&C for the past two years. I find it to be very useful. I would like to see some articles or features on trading plans. I think understanding how to design a trading plan could be valuable for many traders.
- Claudio Castelo Branco, via e-mail BrazilHaving a trading plan is indeed an important part of any trader's discipline. Try some of these past S&C articles on developing a trading plan:
Bryce, James Covington [1989]. "Discipline In Trading," Technical Analysis of STOCKS & COMMODITIES, Volume 7: April.
Chande, Tushar S., Ph.D. [1992]. "Forecasting Tomorrow's Trading Day," Technical Analysis of STOCKS & COMMODITIES, Volume 10: May.
Guppy, Daryl [1999]. "Some Rules To Live By," Technical Analysis of STOCKS & COMMODITIES, Volume 17: April.
Hamilton, Robert J. [1991]. "Managing Emotions And Money," Technical Analysis of STOCKS & COMMODITIES, Volume 9: January.
Hartle, Thom [1991]. "Keeping A Trading Journal," Technical Analysis of STOCKS & COMMODITIES, Volume 9: February.
_____ [1995]. "New Market Wizard Robert Krausz," interview, Technical Analysis of STOCKS & COMMODITIES, Volume 13: September.
_____ [1998]. "Be Rich Or Be Right? CTCR's Courtney Smith," interview, Technical Analysis of STOCKS & COMMODITIES, Volume 16: November.
_____ [1996]. "On The Intuitive Trader: Robert Koppel," interview, Technical Analysis of STOCKS & COMMODITIES, Volume 14: August.
Kaeppel, Jay [1995]. "Should You Trade Futures?", Technical Analysis of STOCKS & COMMODITIES, Volume 13: August.
Sweeney, John [1985]. "Novice Speculator Trading Plan," Technical Analysis of STOCKS & COMMODITIES, Volume 3, February.
Altman, Roger, Ph.D. [1993]. "Designing Trading Systems For The Stock Market," Technical Analysis of STOCKS & COMMODITIES, Volume 11: October.
Arnold, Curtis [1993]. "Developing A Trading System," Technical Analysis of STOCKS & COMMODITIES, Volume 11: August.
Katz, Jeffrey Owen, Ph.D. [1996]. "On Developing Trading Systems," Technical Analysis of STOCKS & COMMODITIES, Volume 14: November.
Morris, Greg [1996]. "Developing A Trading System," Technical Analysis of STOCKS & COMMODITIES, Volume 14: March.
Sweeney, John [1991]. "Developing A System," Technical Analysis of STOCKS & COMMODITIES, Volume 9: November.Past S&C articles can be purchased individually from our Online Store at Traders.com.-Editor
Back to October 2002 Contents