January 2000

Trading Tip:
Fibonacci Time Projections
by Howard Arrington

The Fibonacci sequence involves the use of ratios, and the two ratios that are focused on in this trading tip are 1.618 and 2.618.  Time can be forecast by measuring the horizontal time period between two points A and B, and multiplying by the Fibonacci ratios.  The result is plotted rightward from point A.

Fibonacci time projections can be made from top to top, bottom to bottom, or top to bottom.  Because of the multiplicity of swing tops and bottoms, there is the problem of a multitude of possible relationships.  One solution to the problem is to use this tool in conjunction with other time-oriented information, such as Elliott wave counts or cycle analysis.

The following chart illustrates three Fibonacci Time Projections.  Each time period measured is marked with blue lines and arrows labeled A and B.  The forecasts are marked with red lines and labeled for the Fibonacci ratio used.

Fibonacci Time Projects are easy to make using the Fibonacci Ruler tool in Ensign Windows.   The percentages, colors and markers are configured on the tool's property form.


Study Insight:
Trends Color Bar Study
by Howard Arrington

The Trends study is applied to a chart by clicking on the Color Bars button, and selecting Trends in the list.  Bars in an Up trend are colored with one color, such as green, and bars in a Down trend use a different color, such as red.  This is the logic used to create the Trends Color Bar study.

This code is called in a loop so the test is made for each bar on the chart.  The rPaint array will store a value of 1 to color the bar for an Up trend, or the value of 2 to color the bar for a Down trend.

if rPaint[y]=1 then begin 
  if (r[w].h=r[y].h) and (r[w].l<r[y].l) then begin rPaint[w]:=2; exit; end;
  if (r[w].h>=r[y].h) or (r[w].l>=r[y].l) then begin rPaint[w]:=1; exit; end;
end;
if rPaint[y]=2 then begin
  if (r[w].l=r[y].l) and (r[w].h>r[y].h) then begin rPaint[w]:=1; exit; end;
  if (r[w].l<=r[y].l) or (r[w].h<=r[y].h) then begin rPaint[w]:=2; exit; end;
end;
if r[w].h>r[y].h then rPaint[w]:=1
else if r[w].l<r[y].l then rPaint[w]:=2
else rPaint[w]:=rPaint[y];

r[] is an array of bar records.  The .h field is the bar's high.   The .l field is the bar's low.
The index w references the bar being tested, and y is the index for the adjacent bar to the left.

The programming logic can be expressed with this word description:

If the previous bar's trend is Up, then make two tests:
  If the current high is equal to the previous high and the current low is below the previous low, then color this bar Down.
  If the current high is at or above the previous high or the current low is at or above the previous low, then color this bar Up.

If the previous bar's trend is Down, then make two tests:
  If the current low is equal to the previous low and the current high is above the previous high, then color this bar Up.
  If the current low is at or below the previous low or the current high is at or below the previous high, then color this bar Down.

If the current bar has not be colored by the above tests, then make these tests:
  If the current high is above the previous high, then color the bar Up.
  If the current low is below the previous low, then color the bar Down.
  If the bar's trend is still undetermined, let it be a continuation of the same trend as the previous bar.


Article:
January Effect
by the ChartDr, 12-26-1999
http://www.chartdr.com/

"With the beginning of a new year, century and millennium, investors would do well to reflect on a seasonal pattern to stock movement generally described as the "January Effect". The January Effect describes the physical result of the termination of year end tax loss selling. Looking at seasonal fourth quarter price activity in issues which under performed broad market results during the first three quarters, year in and year out there has been a tendency for lagging issues to perform even worse during the last three months. Tax selling by many institutions in October (timing determined largely by fiscal year end) through December, when public investors tend to offset gains with losses, often serves as a physical weight, depressing share prices below levels justified by companies' fundamental outlooks."

"The natural consequence of this year end selling activity has been described as the oft-anticipated year end or Christmas rally because this concentrated selling activity seems to peak around mid-month. For this reason, every year I like to search for unusual values around this time. A good screening tool at this time is the list of issues making new lows, as such issues may offer extreme values, at least good for a few weeks of resulting gains as tax selling dissipates."

"In fact, such resulting strength may have already begun for a number of well-known issues. The shares of Eastman Kodak, one of the Dow Jones 30 Industrials, rose overnight to 63 3/4 on 12/22 from it's 56 5/8 annual low of 12/21. Since 12/14, the shares of another blue chip issue, Boeing, have risen nearly 10% from a seven month low,37 5/16, to 40 1/4. After declining from 57 1/4 November 18 to 40 1/4 December 20, a three day advance has already carried the shares of Schering Plough to 44 15/16, for a quick 10% gain. Even the interest-sensitive electric utility industry, weighted all year by rising interest rates may be starting to share in this phenomenon. The shares of electric utility Unicom Corp. have risen from 31 5/16, a small fraction above it's 52 weekly low, to 34 3/16 in four days since December 20, again, nearly 10%. Since creating a new 52 weekly low 12/17 at 8 3/8, the shares of Hilton Hotels have risen as high as 9 3/8, for a gain of 12%... so far."

"As we approach Y2K, after reflecting on amazing results of the past 12 months, it appears that the potential for extraordinary gains from the January Effect is enhanced this year as a natural consequence of the tremendous gains experienced this year. Bon appetite!"

(Published by permission of Paul Stockinger.  All views expressed in this article are those of the author.) 


Profile:
Dr. John Arrington

ES:  What is your experience in trading?

JA:  I am a civil engineer, but I have been trading the futures markets for 25+years. I trade Live Cattle and Live Hogs on the CME with an occasional trade in Bellies and grains.

ES:  What do you look for in the markets, and what studies do you use?

JA:  I watch for divergence between prices and the Relative Strength Index indicators for long-term trades.  Time frame for the RSI indicator is not important so I use the same 14-bar Close RSI parameter on all charts.   For day-trades I watch tick charts and 5-minute charts, and just follow my feeling for the market.  I read all the news stories with fundamental statistics, and follow the Cattle On Feed reports and the Daily Meat price reports. So I have an expectation of where the cattle and hog markets should go for the next week.  My short term trades are daily or 1-2 weeks. My long term trades are held for several months.

ES:  I remember in November 1998 you published your cattle research results to the Ensign Discussion Group.  What kind of research do you do?

JA:  That was some interesting cycle research that indicated a high probability of a contract low around the first week of December, followed by a contract high in the middle of March.  The analysis used the last 20 years of data for the April Live Cattle contract.  I do my own research using the ESPL programming language in Ensign Windows, or write my research programs in Delphi.

ES:  What research projects might you be working on now?

JA:  Currently I am interested in Neural Nets.  I am testing short term networks that use daily ticks to predict tomorrow's action.  I am also testing longer term networks that use daily bars to predict the next 10 days.  I use the QNetV2K Neural Net software and have implemented their dynamic link library into my custom programs written in Delphi.  I recently finished automating the data handling process.  Using the ESPL programming language in Ensign Windows and external DLLs, I groom today's data sets, pass data structures to QNet, and plot the QNet forecast as an overlay on the host cattle chart.  The entire process is scheduled in Ensign Windows and runs unattended after the markets close.

ES:  What pit falls do you counsel new traders to avoid as they search for the Holy Grail of trading?

JA:  Don't OVERTRADE.  There are times when all your trades will be against you.  Be ready to take your losses and live to trade another day.  Your First Loss is your Least Loss.  If you cannot sleep at night then reduce your positions to where you can sleep at night.

ES:  Does the Holy Grail exist?

JA:  No, there is no Holy Grail.  There is only lots of work and lots of luck.  Being prepared contributes to the luck that sometimes comes along.  Remember to take a cookie when they pass the cookie plate around once in awhile.

ES:  You have an impressive trading desk set-up.  Can you elaborate on the computer equipment you use and what this set-up cost?

JA:  I have three 19-inch Viewmate monitors I bought last summer for $270 each.  The main screen on the left shows both quotes, charts, and news.  The other two screens show 8 charts each.  My layout has tick, 5-minute, and daily charts for two cattle contracts, and two hog contracts.  I rarely change charts during the day.  My system uses a 300 MHz Hewlett Packard 6330 computer which cost $550 a year ago.  The computer I research Neural Nets with is a 500 MHz Compaq 5600 computer which cost $975 last August.  Both computers run Windows 98 and Internet Explorer 5.0.  Everything is very stable, and truly a dream system compared to what I started with 25 years ago.  I have used Ensign Software products since 1982.


Copyright © 2008 by Ensign Software, Inc.