May 2002

We all seek a leading indicator that gives advance notice of what the market is going to do next.  I have been watching the interaction of JNPR (Juniper Networks) and the June Nasdaq futures contact.  At several key turning points in NQ M2, I see that JNPR is already leading the way, giving advance notice of what the NQ M2 market will do next.   Permit me to share with you a three day sequence in comparing both charts using 5-minute intra-day bars.

On Monday, NQ M2 formed a wedge from the open until 12:00 (noon Central time).  The question in the minds of NQ traders was whether the breakout from the wedge would be up or down.  The JNPR chart was steadily lower all morning serving as an advance indicator that NQ would be going DOWN and not up.

On Tuesday, JNPR and NQ M2 both rallied until 10:30.  NQ M2 put in a new high at 11:05, but JNPR did not.  This is a first BIG clue that the 11:05 top in NQ M2 was going to be the daily high.  At 14:10, NQ M2 was pushing higher for a possible retest of the 11:05 high, but JNPR was definitely in a downtrend.  JNPR's 14:10 rally was nowhere near its mid-morning high.  Again, JNPR served as an advance indicator that the NQ M2 retest of the high would fail, and NQ M2 would go lower.

Wednesday opened with a sharp sell-off in both JNPR and NQ M2.  They both put in lows at 9:35.  NQ M2 put in a new lower low at 10:30, but JNPR found support and did not put in a new low.  Again, JNPR gave advance notice that the market would not continue lower.  See the divergence that is marked on this chart.  From the 10:30 lows, both JNPR and NQ M2 experienced significant rallies.

Have I found the crystal ball for NQ in the form of JNPR?   I doubt it.  Useful correlation found (with the benefit of hindsight) on three charts is not statistically sound.  However, the trading tip being taught is that an individual stock may respond a bit quicker than a composite index.  JNPR is a heavily traded Nasdaq stock, and I am seeing useful correlation where the e-mini Nasdaq future follows the direction set by JNPR.  At least in the three day sequence documented in this article, JNPR was a worthy leading indicator.  There was significant divergence at three key turning points in comparing JNPR behavior at these tops and bottoms with the NQ M2 behavior.  NQ M2 followed the JNPR lead either higher or lower.  When NQ M2 tried to go higher but JNPR was going lower, NQ M2 turned around and went lower.  When NQ M2 tried to go lower, but JNPR would not follow lower, NQ M2 turned around and went higher.  This is just the type of leading divergence indicator sought after by analysts who use studies to show confirmation and divergence.

Appreciation is expressed to Greg Corbin for sharing this amazing chart of July Wheat showing the Gann Square of Nine levels marked on the chart.  Note how frequently the market turns on one of the 45 degree intervals on the Gann Square of Nine.

The Square of Nine is a spiral of numbers on a grid, starting with 1 in the center.  The spiral gets its name from the first inner square counting to nine.   The left side horizontal is known as the zero or 360 degree line.   90 degrees would be the prices that align vertically above the center point, and is measured clockwise from the zero line.

Example beginning of the Square of Nine number spiral:

31  32  33  34  35  36  37
30  13  14  15  16  17  38
29  12   3   4   5  18  39
28  11   2   1   6  19  40
27  10   9   8   7  20  41
26  25  24  23  22  21  42
49  48  47  46  45  44  43

The number series continues to spiral outward to ever increasing values.   Eventually the values shown on the chart are discovered on the Gann Square of Nine.  On the Square of Nine the number 2650 is on the 180 degree ray, and the number 2980 is on the 360 degree ray.   The Gann Square of Nine is a tool available in the Ensign Windows program.

Part I

Does history repeat itself?  Is there something to MOON PHASES, TRIPLE WITCHING, LUNAR and SOLAR ECLIPSES?  What about FOMC DAYS, or WEEKS, or any other possible repetitive/cyclical possibilities you can think of?  What of PATTERN RECOGNITION, and PROGRAM TRADING?  Is there really anything to all this stuff, or is just another bunch of hocus pocus?

Did you know that, since 1982 (the first year the SP FUTURES started trading), FULL MOON TRADING DAYS in AUG close NEGATIVE 76% of the time?  Here are the only times in SP history that the FULL MOON was on a Thursday in August.

Tip:  This year the FULL MOON is on Thursday, August 22, 2002.

I've invested in the markets for some years now.  At one point I was getting a little bored.  So a friend of mine suggested that I check out a website www.programtrading.com.  It was back in 1997 and I read all this stuff about when the CRASH was coming.  The site wrote about some Monday in October, and that it was the Anniversary of the Crash of 29, that it was some color and some moon thing and some other stuff.  I said to myself, "sure".  October came and guess what?  Yes, we did have a market drop of over 7%.  And Yes, it was on the date that was mentioned on the website.  So I decided to check out one of these Program Trading Seminars.

This is the 5 minute chart of the "CRASH OF '97"

This seminar changed my whole view on things.  I proceeded to gather all types of historical data, and created a database of many things.  There was a problem however, in that I couldn't get the historical charts to look at all like what I am used to looking at every day.  I spent some time developing this database, but yet could not apply it in a simple way.  The simplest way for me would be to be able to print out historical charts that would match my existing real-time software. I've used Ensign Software since 1997, and I have found that there is no better software in the world for me.  It's so easy to use and so easy to SEE.

So I set out to create a charting software for historical data that could do what I wanted it to do.  A software to give me the flexibility for viewing and printing that I needed.  AND one that had no limitations with regard to sizes of data files.  Now with this software I can print out any charts I want and it is if I just printed it from today's trading day.  I found this to be Very Powerful.  It's like having duplicates of the same photo.

Is it real, or is it just a random walk?

Expiration Fridays in December close POSITIVE 80% since 1982.  That's 16 of 20.  Knowing this, wouldn't you want to have a printed chart of every one of them, in your preferred time frame, and it looks just like what you are looking at real time?

These are 2 of the December Expiration Fridays. GDP report was also released these days.  Yes, it was the only times that the GDP was released on EXP Friday in December.

Tip:  I believe that this year, Expiration Friday is 12/20/2002.  And I believe that GDP will be released the same day.

How about this?  When Expiration Friday in December is on the 15th of the month, (3 times in SP history), there were 3 of the 4 times that a December Expiration Friday closed NEGATIVE.  The HIGH of the morning was 8:35 am.  Wouldn't it be nice to have those charts printed out and at your fingertips?

I was on a plane one day with a friend of mine.  We were on our way to another one of these seminars.   While we were waiting, I started doing some research on that particular trading day.  I said, "Hey Tom, look at this.  Almost every one of these days sell off after the first hour of trading.  Let's call this in".  So I called the broker from the plane, and instructed him to sell some S&Ps at around 9:30, and put a stop above the high of the day.  Then I told him to tighten stops at lunch, and look for the lows of the day to be in the last hour.  He complied.  When we got off the plane and got all our things together we gave a call to see how we made out.

That day was April 14,2000.  The high of the day was 9:15 and the low of the day was 2:50.  The S&P dropped 88 points that day.  Needless to say, we were quite pleased.

Here's an interesting one.  When the First Trading Day of JULY is on a MONDAY, the SP'S have closed POSITIVE 100% of the time.  These are just a few of those days.

(Editor's note:  Dom's article is lengthy, but very intriguing.   The balance of his article will be published in the June 2002 issue of the Trading Tips newsletter.)

Part II

In the first part of this series, I covered some basic concepts about ESPL.  This segment will examine ESPL for essential characteristics of a programming language.

Over the past few years, I have looked at, and used a small hand full of programmable charting packages.  As a programmer of many years, having worked with over a dozen programming languages, the first thing I look for in these charting packages, are the essentials that are key to any programming language.  I am happy to say that ESPL has the largest number of these essentials.

Before other experienced software developers hand me my head, let me point out that there are several different types of programming languages.  Two classifications that come to mind, are what I call, "linear programming" and "rules based programming".  In linear programming, the program is executed in the order that the code is written, unless modified by flow control instructions.  Conventional linear languages are C++ and Java.  Rules based languages can execute in what might appear to be random order, based on criteria being met, regarding the data being operated on.  Examples of such languages are Amzi Prolog and XSLT.  ESPL is a linear programming language.

Standard Programming Language Features

The first thing presented in many manuals and books for the typical linear programming language, is a description of the data types and variables the language supports.  Most coding in ESPL can be done with the data types:  integer, real, and Boolean.  In addition, ESPL provides support for arrays of these basic types and character strings.  There are also several "special" data types, that are unique to ESPL, such as TDateTime, TListBox, TStringList, TFont, TForm and TScreen.  These special data types are used by specific built-in methods.  As expected, your code does not access their contents directly, but rather obtains ones of these data types from a built-in method and then passes it back to other built-in methods.

The next thing typically discussed in programming language manuals is operators and expressions.  In this area, ESPL has the full complement of numeric, Boolean, and string operators.  The numeric operators include binary operators such as shift left (SHL), shift right (SHR), binary OR (OR), binary AND (AND), and exclusive-or (XOR).  All of the typical Boolean operators are provided.  For strings, the ‘+’ operator provides concatenation.

Flow control statements are next on my hit list.  These include the typical goto, if-then-else, while, repeat-until, and for.  ESPL has all of these statements.  I include in the flow control group, what I call convenience statements.  These would be break, continue, and return.  Unfortunately ESPL does not have these statements.  Flow control also includes the ability to call subroutines and functions.  ESPL uses the name procedure for subroutines, and function for functions. A built-in variable called Result must be assigned the return value of a function, and return is made at the end of the function’s code.  The lack of a return statement leads to the use of a label, placed at the end of the function, with goto statements in the function’s code.  I guess I should have mentioned that ESPL has the label data type.  Labels must be explicitly declared.  Note that a declared label that is not used, both as being placed in the code (i.e. label: ), and being referenced (in a goto statement) will result in a compilation error.

The mention of procedures and functions requires a brief discussion of parameter passing.  Conventionally there are three ways to pass parameters to a method. (The word method is used here as a generic term for procedures and functions.)  They are referred to as: pass by value, pass by reference and pass by name. Pass by name is relatively rare, and does not appear in ESPL, so we’ll skip that one.  In the case of pass by value the method obtains the value of the parameter, and can not effect any variable in the calling method.  With pass by reference, the called method has access to the value of the parameter, and any change made to the parameter will change the associated variable in the calling method.  The primary use for pass by reference, is to allow the called method to provide multiple outputs to the caller.  This is in contrast to the single output provided by the returned value of a function.  Happily, ESPL provides for both pass by value and pass by reference.

A programming language is of no use, without input and output (usually referred to as I/O).  Here is where ESPL truly excels.  Most of ESPL’s I/O deals with interfacing to Ensign charts, which I will discuss in a moment.  In addition, ESPL has a full complement of other I/O methods. These include:

• the reading and writing of files;
• reading keyboard input;
• writing to the text output window of the ESPL editor;
• writing to a text box window (in the main Ensign window);
• writing to the Windows clipboard;
• reading and writing to DDE (Dynamic Data Exchange);
• playing sounds (.wav files);
• invoking Windows/Dos commands;
• reading http sites from the internet;
• reading and writing ftp sites from the internet; and
• sending e-mail.

You must admit, that such a list of built-in I/O capabilities is very impressive for any programming language.

An additional capability, which could loosely be classified as I/O, is the ability for ESPL to invoke any code made available through a DLL (dynamic link library). With this capability, an ESPL script can do just about anything.

The last conventional programming language feature, that I want to mention about ESPL, is the Include statement.  This statement allows the programmer to segment the code being written, into logical pieces, each in its own file.  Those files can then be "included" into a master file, which is the file opened in the ESPL editor.  When the program being written is large, the capability to segment pieces of it, is essential.  Though this statement has nothing to do with execution, it is (in my humble opinion) a critical feature in any programming language.  I am happy to say, that, with regards to ESPL, it’s in there.

Charting Package Programming Language Features

A general programming language is all well and good, but we are interested in a charting package programming language.  Here is where ESPL will blow you away.  To state our desire in one sentence: anything I can see on the screen (charts, quote page, etc.) and anything I can do with the mouse and/or keyboard (apply studies, etc.) should be accessible to the ESPL program.  I would say that ESPL meets this goal about 90%.  Far greater than any other charting package programming language that I have seen thus far.

Starting with a basic chart, ESPL has full access to the values of the price bars.  This includes the obvious High, Low, Open, Close, Volume, and Timestamp. ESPL can read and write these values for any bar on the chart.  In addition, ESPL can add and delete bars from the chart, thus a program can construct it’s own chart full of bars.

Any study, bar coloring, or draw tool that can be added to a chart, from the menus, can be added and removed, from ESPL.  In addition, any value, for any line drawn by the studies, can be read.  All properties (with a few minor exceptions) including colors, line style, markers, and marker colors can be read and changed from ESPL.

The contents of quote page cells can be read from ESPL.  I have had some problems writing quote page cells, though the documentation says it is possible.

ESPL has built-in methods for activating many of the pull down menu selections.  If that isn’t enough, ESPL can "force" keystrokes into the input stream, as if they had been typed at the keyboard.  An ESPL script can access the list of open windows, it can open new windows, it can close any window, and obtain information about any window – such as its contents, if it minimized, etc.

I hope you are beginning to see that the power of ESPL is awesome; and the possibilities for what it can do, are endless.

For reasons of both size and subject matter, this is a good place to stop.  Next time, we will look at the issue of data storage for your programs, and then several new features that I would like to see in future releases of ESPL.