Notice:
  The views and information expressed in this document reflect the opinions and experience of the author Robert C. Pelletier.  Neither CSI nor the author undertake or intend to provide tax advice or trading advice in any market or endorse any outside individual or firm.  All recommendations are provided for their informational value only.  Readers should consult competent financial advisors or outside counsel before making any software purchase or investment decision.  CSI does not stand behind or endorse the products of any outside firms.

Copyright (c) 1998 Commodity Systems Inc. (CSI).  All rights are reserved.

OVERVIEW

   December leaves us with a hodgepodge of topics that have emerged throughout the year, but have not been addressed in recent Technical Journals. This month we will discuss Master File Directories, Y2K compliance for the MetaStock® Format, Day-only trading sessions, Perpetual Contract Data and Understanding your database.

Topics discussed in this month's journal.
 

Why "Compatible" Software May Not Be  Y2K and the MetaStock Format The Sun Always Shines Somewhere Perpetual Contract Data and Market Volatility Understanding Your Database Identifying the Last Trading Day Questioning Fed Funds Correction to EMU Action Table

   One of the great things about Unfair Advantage® is that it is almost universally compatible. It can produce data files that can be read, in one form or another, by almost every trading program on the market. This was accomplished by programming the specific requirements of the data formats into our software and giving the user the option of selecting these formats for file creation. QuickTrieve® also offers compatibility with many programs in the same way. One thing Unfair Advantage and QuickTrieve can't do is correct programming deficiencies in third-party programs that do not adhere to these formats. Failure to comply with the record layout may cause the third-party software to access the wrong financial time series from your CSI database.
In the case of the CSI (QuickTrieve) format, the accessing software must read the Master File Directory to identify the symbol, name, CSI number, stock or commodity flag, delivery year and month and other details about the files held in the directory. The Master File Directory identifies the 'F' file number (such as F001.DTA) for each data file held in the directory, allowing the accessing software to address the desired file.  This system makes it absolutely impossible to address the wrong stock or commodity contract when the Master File Directory is used properly.
Despite clear descriptions in our published record layout, some software developers have consistently ignored the Master File Directory when attempting to locate files in a given directory. Some incorrectly assume that the 'F' file number will remain the same from day to day and rely on that number, (not the Master File Directory) to find the issue. The 'F' file number can easily change because of deletions and additions of new commodity contracts or stocks. Files come and go from a directory because of the nature of the birth-death process of the commodity contracts, merging of stocks, etc. Failure to read the Master File Directory can result in improperly identified files.
The Master File Directory provides two symbol fields for the commodity sought. One holds a two-character symbol, which was sufficient several years ago when fewer commodities were traded, and the other provides for a three- to six-character symbol. This will accommodate the entire symbol universe for all commodities and stocks traded today. Some software developers have pretty much refused to address any symbol length of three characters or more. When trading software looks at only the first two characters of a symbol, confusion can result. For example, ED2 (Day-only Eurodollar) can truncate to ED (combined-session Eurodollar). Can you tell which is which?

Y2K and the MetaStock Format

   CSI's format is Y2K compliant, which is one reason for its enduring popularity. MetaStock's format will probably be Y2K compliant, but so far as we know, their format has not yet been publicly disclosed. We have been advised by Equis that they will not disclose to CSI the MetaStock format details that would allow us to carry this format forward into the 21st century. This does not mean that we will not support the new MetaStock format; it just means that we haven't been advised as to how it is constructed. It may take some time to determine what should be done to accommodate our users and/or work out some alternatives.
On the other hand, MetaStock may have some import features or other procedures that might make it possible for you to avoid discontinuing use of your MetaStock software or settling for a data service that may not have all the information you need. If you use the MetaStock format produced by CSI, but you do not use MetaStock software, then there is no need to be concerned. We will force the old format to be compliant so that it will still work with MetaStock-compatible software. 
The company that makes MetaStock is owned by Reuters®, which also vends end-of-day data and considers CSI to be a competitor. MetaStock supports other competitive-to-Reuters data vendors that are not owned by Reuters, but for reasons known only to them they are not considered to be a competitive threat. Nevertheless, many MetaStock users enjoy the benefits of using CSI data with their analysis software. Unfortunately, CSI users who depend upon MetaStock software may be caught off guard when January, 2000 arrives. If you would like to advise the makers of MetaStock of your views, you might consider writing to Bryan Strain (bstrain@equis.com) at Equis.  Perhaps he can offer some help or advice. In the meantime, we are doing an informal survey of MetaStock users. If you use MetaStock software for analysis, please send us an email describing your data usage to Techsupport@csidata.com. Please include the words "MetaStock Survey" in the subject line

   Since it's always daylight somewhere (and some traders just don't sleep), the demand for 24 hour markets keeps growing. We have promised to address the oft-confusing issue of after-hours trading, and which hours are included in the various contracts. The following table should answer some of your questions.

   Perpetual Contract data is a proprietary CSI service that synthesizes the prices of many contracts into a single data series. The whole idea of Perpetual Contract data is to monitor price movement where most traders are in the market. Those who advocate using back-adjusted market synthesis may be too close to market action, where it is typically more volatile. Other computed series such as Gann and Nth nearest futures with larger values of N may have an excessively distant view of the market.
The Perpetual Contract data series that does an especially good job is the one which weights prices by volume and open interest (code 51). The 49 Perpetual is also a good choice because it rolls at about the right time and is sufficiently distant that all markets can be treated the same.  Perpetual Contract data fashions a computed slice in time at the right distance from a market to monitor activity where traders actually trade. If you base decisions on the full history of a contract using other means, you may be missing the action where volume and open interest intensity is the greatest. Granted, the Perpetual can't be traded per se, but it can be an excellent barometer to tell you what to do in the near contract where your position is applied.

What Customer Mark Brown is Saying About Perpetual Contract Data.

   "I use the Perpetual Contracts [data] myself and am surely a minority. When a future contract nears expiration it becomes more volatile. That volatility causes many systems to initiate trades that would have otherwise not been engaged. If the real market is to be traded, surely then it would exist somewhere in between the current contract and next one out. Given that, I would suggest you think about this: Many people build models using back adjusted contracts and many other methods. All of my current EOD models work best on that type of data. However when you understand why they work better then you may want to reexamine the way you build models. It sounds like you have a model that follows price action. I propose that a good model instead trades a markets historical 70 to 90 percentile personality. The very moves that others concentrate so hard on are the very moves I dismiss as an anomaly. mb".
 
 
Understanding Your Database

   Among the variables we deal with at CSI are the naming conventions each exchange has either developed or adopted as a way of identifying contract delivery months. On the London Metal Exchange, for example, contracts are identified based on the explicit number of months or days forward from the current date they will become deliverable. Such markets are called "forward markets." At U.S. exchanges and many other world exchanges, futures contracts are defined simply with a particular calendar month of a given year, such as March, 1999. This convention gets somewhat confusing when an exchange identifies a delivery month that is different from the actual delivery month. March contracts, for example, have been known to expire during the month of February.
 
 
Identifying the Last Trading Day

   At CSI, we have gone to a great deal of effort to define the contracts and commodities that are traded worldwide, and as a convenience, we often show a targeted last trading day for each commodity. The last trading day information can be especially valuable when dealing with a contract that expires before its "delivery" month. In actuality, the last trading day of any given commodity is based upon an exchange rule that would force the date to be exact, such as the third Tuesday of the month or the first Thursday prior to the month end, etc. We have translated that convention into a more or less arbitrary date, which is stored in Unfair Advantage's CDBFACTS.ADM file and is visible through the pull-down fact sheet from the tool bar.
Keeping in mind that the date is not necessarily exact, you can find the actual month of delivery by checking the "Last Trade Date." If the value shown is less than or equal to 31, then the specified "delivery month" is the actual expiration month. If, however, the last trade date is specified as a quantity greater than 31 (up to 62) then the last trade date will occur in the month before the specified delivery month. The date will be the specified value less 31. This exception covers the case where the exchange stops trading on the month before the specified delivery month. If the CSI specified last trade date is 62, for example, then the last trading date would be the 31st of the month prior to the exchange identified delivery month. If the last trade date is not specified for a market of interest to you, please contact CSI and ask for identification in our printed or on-line fact sheets. We will make every effort to accommodate your request.
 
 
Questioning Fed Funds

   Sometimes even the most meticulously kept database can appear erratic due to the nature of the data. For example, our Fed Funds series is as close to being perfect as any series we offer, but a quick look at the data (CSI commodity numbers 74, 172 (code 54) and 398) may make a user question the accuracy. CSI's data series, which are supplied by the U.S. government, reveal huge distortions that often coincide with the end of each calendar quarter. These fluctuations merely reflect the US Federal Reserve's (Greenspan and the Federal Reserve Open Market Committee) efforts over the years to use the Fed Funds rate to control the money supply and keep inflation forces under control. 
Interest rate adjustment action could be a warning sign that might help you in your overall decision making. Be assured, it is not likely to be a bad CSI data point. We researched all such suspicious jumps and drops in the rates over the decades and made any necessary corrections. For more information on just how the Federal Reserve works, please see past CSI Technical Journals from Aug. & Sept. '96.
Unfair Advantage offers the unprecedented ability to perform in-depth studies on all North American markets or on all markets around the world.  Now the task is yours to understand the data you have at your fingertips.  Along with the information we have been publishing in recent Journals and in our fact sheets and manuals, our new web site (http://www.csidata.com/) offers links to exchanges around the world. We urge you to use any and all of these sources to answer questions you may have. We want to help you get the most from your database, so that you can make the best choices for your investment dollars.
 
 
Correction to EMU Action Table

   Our November newsletter included an EMU action table advising commercial customers and users of QuickTrieve how they might adjust their affected data series to accommodate the upcoming conversions. We regret that the table included an error regarding the adjustment of volume and open interest figures.
Reader Kirk Ewald noticed that our advice did not produce pre-conversion day volume and open interest accurately.  Upon closer examination we recognized our error:  The old volume and open interest readings should actually be replaced by the ratio of the new Euro contract size times the conversion factor of the old currency (as the numerator) and the old currency contract size (as the denominator).
In the case of CSI#132, the German Long Bund, for example, the new Euro contract size is 100,000 Euros and the new Euro would be worth approximately 1.968 D. Marks if we used the conversion rates effective on September 30, 1998. The pre-conversion day contract size is 250,000 D. Marks.  Therefore the ratio would be (100,000 Euros * 1.968 D. Marks per Euro)/250,000 D. Marks.  The numerator becomes 196,800 D. Marks because the Euros cancel out.  The denominator is already in D. Marks, so the ratio becomes the factor which should be multiplied by the historical volume and open interest readings. In this example, the multiplier would be .7872.  The end result should be rounded up by one half a contract to account for the convention of holding volume and open interest readings as an integer.
If you apply the same logic to the Indices such the DAX Index in Germany, you would have the new index multiplier of 25 Euros times the index for the new Euro contract to be equated with 100 D. Marks times the index for pre-conversion days.  Converting all units to D. Marks again, you would have a new ratio of 25 Euros times 1.968 D. Marks per Euro for the numerator and 100 D. Marks as the denominator.  The Euros cancel out in the numerator again, producing a ratio of (25*1.968D.Marks)/100 D. Marks, or a multiplier of .492 (after canceling out the D. Marks) to be multiplied times historical volume and open interest readings in pre-conversion day history. This would be the case in spite of the left-over units on conversion day because fractional units between old and new index multipliers that don't exactly agree with the conversion rate are settled in cash or as a one-time security issue on conversion day.
 
Please note:

   Unfair Advantage will automatically compute these changes using the above formula.