Ask The Analyst, Sunday, 04/25/2004
Tendency to gravitate toward the pivot (Pivot Analysis) By Jeff Bailey
This week's question comes from a stock index swing trader, where the question was a complex one, and one I've really had to work at.
The questions, which were quite complex, involves back testing,
where some of our historical data from the weekly pivot analysis
matrix might help us all get a better sense of how these markets
have been trading, while at the same time, give us some insight
as to what our expectations for profit, or price action might be
over the period of a week or two, even three or four weeks.
I love questions that I don't immediately have the answer to, and
if I feel it is within my mental capacity to answer, I love a
good challenge.
The swing trader's question centered around his observation that
on a weektoweek basis, the major indices seemed to gravitate
toward, or around their weekly pivots, where even when
fluctuations away from the pivot were found, there still seemed
to be some type of gravitational pull, where swing trade profits
could easily evaporate by the end of the week.
Trader's Question #1: Is it just me, or is this really
happening?
Trader's Question #2: What type of range, or fluctuation is
normal for the QQQ? It seems that there are periods of
volatility that offer opportunity to profit, but many times, that
opportunity must be captured quickly, or the gravitation back
toward the pivot suddenly has paper profits suddenly missed.
What type of swing trade profit on a weektoweek basis should I
look for. Maybe my expectations are too high, thus some of my
frustration.
Trader's Question #3: After following the QQQ's trade on a week
toweek basis, it been my observation that some week's volatility
or price fluctuation is much greater than others, but as you
noted in last night's wrap, the QQQ has become a much more
difficult security to figure out, as it is now blended with
stocks other than pure technology related stocks. Of late, the
QQQ hasn't necessarily been trading as much in unison with the
semiconductors as it had in the past. Am I seeing things, or
have the semiconductors lost some of their usefulness when trying
to determine price direction of the QQQ?
Wow! I could spout off some anecdotal answers to these three
question, but that's no good.
After thinking about this trader's questions for a few days, I
thought I run a little experiment. Much like a scientist will
do, I'll use some past data, make some observations, and see if
we can't come up with some conclusions.
Here's what I came up with.
To get a grasp of what I was dealing with, I went back and
compiled the last 12weeks of WEEKLY pivot analysis data. I keep
track of all the weekly pivot data I show in each night's Index
Trader Wrap at OptionInvestor.net, so we can use actual data to
begin addressing the trader's questions.
The last 12weeks data should give us a large enough data set for
a swing trader to begin making some useful observations from.
The periods I used are from February 2, 2004 to Friday's close of
April 23, 2004.
Here's how I hope to tackle some of the trader's question, and
while I was at it, I included all of the various
indices/securities we keep track of in the pivot analysis matrix.
Data Table of Weekly Pivot Analysis 
Here's how I went about tackling some of the trader's questions.
Stick with me, and I think it will become clear, but with the
resources I have at hand here's what I've done.
The only way I could figure out how to display each week's trade,
was to assign it a number. Since I'm using 12 weeks of data, I
assigned each week a number of 1 thru 12. This is displayed at
the bottom of the above table. As an example, for the week of
2/2 thru 2/6, I've assigned that week's trade as #1, then indexed
up from there. Simple enough.
Now comes the hardest part. I went back and looked at what the
price was at each Friday's close for the particular week (1 thru
12) and then plotted that each Friday's close in the above table
(rows 414) for the appropriate index/security we track in the
WEEKLY Pivot Analysis table. Each week's closing value was
plotted in relation to where the index/security closed with THAT
WEEK'S weekly Pivot Analysis Matrix.
For the purposes of this article, and as to try an make this
article a broader educational on a broader market perspective, I
used the S&P 500 Index (SPX.X) as my general market indicator of
price direction. Still, a NASDAQ100 (NDX/qqq), Dow Industrials
(INDU/dia), S&P 100 (OEX), Semiconductor (SOX.X), or Banks
(BIX.X) trader will be able to make some observations that they
are interested in.
Let's walk through the process of plotting the SPX (row 6)
closing values for some of it's Friday closes. When we do so,
consider the vertical blue line, which divides column D and E,
the WEEKLY Pivot, which would be viewed as a potential midpoint
of the week's trade. After the completion of each week's trade,
we calculate a NEW weekly pivot for the next week's trade. As
such, we KNOW that when the new week of trade begins (Monday)
that the index/security we are looking to trade, will likely open
for trading either side of the pivot. For example, the SPX will
either open for trade above its Pivot and below R1 (resistance 1)
or will open below its Pivot and above its S1 (support 1).
At the markets close of 02/06 (#1), the SPX (row 6) CLOSED
between is weekly and WEEKLY R1 for that week's (for 02/0202/06)
trade. I plot the #1 in row 6/column E. This gives me
perspective of where the SPX CLOSED relative to a midpoint. Take
special note that I stress CLOSED at a price between the Pivot
and R1. During any given week, the SPX could have traded any of
the levels in the WEEKLY Pivot Matrix, or no levels at all, as if
stagnate.
Now we progress to the week of 02/0902/13 (#2), where that week,
the SPX closed ABOVE its weekly pivot. I plot the #2 in row
6/column E. (Note: SPX traded as high as WEEKLY R2 of 1,155.01
+16 points this week, but never trade AT its WEEKLY Pivot, but
closed up just 3.05 points for the entire week).
We progress further to the week of 02/1602/20 (#3), where that
week, the SPX closed between its WEEKLY Pivot and S1. (Note:
SPX traded as high as its WEEKLY R1 of 1,156.90 +11 points this
week, but closed down 1.7 points for the entire week).
Progressing further to the week of 02/2302/27 (#4), we see that
the SPX once again closed between its WEEKLY Pivot and S1. I
plot the #4 in row 6/column D.
For purposes of brevity, I continued this process for the
remaining weeks of trade in an attempt to try and answer the
trader's question #1.
Answer to #1: After plotting each week's CLOSING trade, it does
appear that CLOSING values, on average, tend to gravitate around
the pivot, or at least between their WEEKLY S1 and WEEKLY R1,
with Pivot in between. Based on 12 weekly observations, this
would be true 75% of the time (9 out of 12 weeks).
Trader's Question #2: I went back and calculated the AVERAGE
weekly range of trade for the various indices/securities and
posted the average range of trade (from each week's low to high)
for the last 12 weeks.
Answer to #2: Row 10/Column H would show that the AVERAGE range
of trade for the QQQ is $1.35. A swing trader that holds a
position no more than 5 trading days, would begin to understand
that if they entered a QQQ trade at the PEAK inflection point
(high or low) of that week, then they might look for a $1.35 per
share gain on average. I don't know any traders that have been
able to consistently buy and sell extreme peaks or inflection
points in ANY security, but traders that can enter and exit
within 15% of an inflection point on consistent basis would be
considered a very good trader. In the context of $1.35, if
missing the outlying 15% of $1.35 range on both end, this would
still have the trader capturing 70% of the average range, or
roughly $0.95 per share in 5days. Perhaps a swing trader that
only prefers capital exposure of 5days or less, but has been
targeting a $2.00 per share gain in any one direction, begins to
understand that this type of opportunity would be outside the
average weekly price fluctuation of the QQQ.
Trader's Question #3: Has the QQQ and the SOX.X lost some of
their correlativeness?
To answer this, I'm going to separate out the last 12 weeks of
trade (1 thru 12) in for both the QQQ and the SOX.X, and plot
each week's close, within each security's weekly pivot matrix.
QQQ and SOX.X weekly closes within their WEEKLY Pivot Matrix
With no regard to the actual amount of price gain or decline,
I've plotted the 12 weekly (1 thru 12) closes of both the QQQ and
SOX.X against each other, where RED numbers would depict a
certain price decline, numbers is BLUE would most likely depict a
more fraction gain/loss or relative unchanged price, while GREEN
numbers would depict a certain price gain.
There does appear to be some lack of correlation, at times,
between the QQQ and SOX.X.
Note that in week #2, the QQQ appears to have been relatively
unchanged, while the SOX.X would have closed below its WEEKLY
Pivot, after having closed ABOVE the prior week's (#1) pivot.
However, I will note that both the QQQ and SOX finished
lower(QQQ= $0.19, SOX= 8.84) by their close for the week of
02/0902/13.
Week #3, the QQQ would have closed BELOW its weekly pivot (fell
$0.08 this week, but did trade as low as its WEEKLY S1 of $36.57
before closing at $36.86) while the SOX.X closed below its WEEKLY
pivot for a secondconsecutive week (fell 0.77 points, no trade
at its WEEKLY S1).
Week #4 and #5 would most likely have shown fractional Friday to
Friday closing changes. I checked the full RANGE of trade for
week #5, which did show the QQQ's weekly range being $0.97, while
the SOX.X witnessed a 22.71point range.
Week #6, shows that a notable downside move (to the left of
pivot) was witnessed as both the QQQ and SOX.X closed BELOW their
WEEKLY S2s. While the QQQ closed down $1.12, or 3.06%, its
weekly range was $2.02. The SOX.X closed down 19.14 points, and
its weekly range was 36.21 points.
Week #7 was a certain price decline. Remember, that when each
new week of trade begins, we KNOW that the index/security being
measured in the pivot matrix resides somewhere between its S1 and
R1, so a close between S2S1 would have to represent a decline.
This week, the QQQ saw a $0.76, or 2.14% decline and traded a
range of $0.99, while the SOX witnessed a 21.75 point decline, or
4.48% loss, and saw a weekly range of 26.28 points.
Week #8 may have been a marginal bullish week, but with a certain
price advance. The QQQ gained $0.47 on the week (+1.35%) but saw
a range of $1.58 (traded down to WEEKLY S2 of $34.01 with a
weekly low of $34.00 before gravitating back above its WEEKLY
Pivot) The SOX gained 15.90 points (+3.43%), and similar to the
QQQ saw a wider range of 34.39.
Week #9 saw prior week's gains build with the QQQ gaining $1.88
(+5.34%) while its weekly range was only $1.66. How can this be
you might wonder? The QQQ started the week with a gap higher
from its prior Friday's close. The SOX gained 34.61 points
(+7.22%) and saw a weekly range of 35.27 points.
It is weeks #11 and #12 where the trader begins to pick up on
some recent lack of correlation between the QQQ and the SOX.X.
Week #11 does find the QQQ and SOX.X moving lower, but the SOX.X
outpacing a decline to close BELOW its Weekly S2, while the QQQ
closes between its WEEKLY S1 and Pivot. This week, the QQQ
finished down $0.85, or 2.3% and traded a range of $1.57. The
SOX.X finished down 31.64 points, or 6.18%, and traded a range of
44.09 points.
Week #11 would either depict that there was something overly
"wrong/bearish" taking place in the SOX.X, or there was a few
stocks, nonsemiconductor, that were showing more bullishness
among the QQQ.
Week #12 would have found the QQQ with certain gain and this
week's trade did find the QQQ closing up $1.12, or 3.10%, with a
weekly range of $1.63. The SOX.X closed up 7.84, or 1.63%, and
saw a ranged of 22.3%.
Once again, we may note that either the SOX.X is seeing something
"wrong/bearish" taking place, or there was a few stocks (I can
think of MSFT's 6.12% gain on Friday) showing more bullishness
among the QQQ.
Answer #3: It would appear that some of the QQQ/SOX.X
correlation has been lost, but only in the last two weeks.
To be continued:
Good gravy! It's 05:34 AM EDT and I'm still working on this
article.
I think I've "discovered" some things and I want to work on them
some more.
I will be on vacation all next week, and won't get a chance to
continue my train of thought until May 9th Ask the Analyst, but
will give me an opportunity to test a couple of things I think
that I've discovered.
Try giving yourself a rest right now, clear your mind, then come
back and study, or reread this article.
Try applying a simple IF the index does this, then trade this
level, with an expected average range of $x.xx, with a stop x% of
the entry point.
Since I will be out all next week, I want to post this upcoming
and now updated WEEKLY pivot levels for 04/264/30, along with
Monday's DAILY pivot levels, and current MONTHLY Pivot levels.
Pivot Analysis Matrix 
Here's what I really want to check against when I get back from
vacation, but something you might want to do this week while I'm
gone.
Think about the AVERAGE range of trade we calculated in the first
chart (column H). IF the average range for the QQQ is $1.35 over
the past 12 weeks, this might have the QQQ capable of trading
($37.21 +/ $1.35) a MAX LOW on AVERAGE of $35.86 (that would be
just below this WEEK's S1, or capable of trading a MAX HIGH on
AVERAGE of $38.56, which would put the QQQ above its WEEKLY R2.
Here is something I'm going to test when I get back.
I didn't mark this correlative level in the Matrix, but do you
see MONTHLY R1 at $37.34 and DAILY R1 of $37.35? Think of this
as a KEY LEVEL OF RESISTANCE right now, where without
help/strength from the SOX.X, it may be difficult for the QQQ to
get above this level.
After all, we KNOW that the QQQ traded this $37.34 level earlier
this MONTH (MONTHLY R1 was traded BLUE UNDERLINED) and look where
the QQQ LOW was this week ($35.63).
On Monday, my thinking would be this, and this is a test. IF the
QQQ trades above $37.34$37.35, then go long (look for some
strength in SOX.X) on break higher, STOP under DAILY Pivot, with
a MAX WEEKLY AVERAGE gain target of $38.56. If QQQ trades WEEKLY
R2 of $38.33, immediately raise stop to protect against
gravitation back to the WEEKLY Pivot.
The BEARISH traders looks for this.
IF the QQQ trades $37.34$37.35, but SOX.X lagging, then be
patient, and should SOX.X begin to weaken the GO SHORT the QQQ on
break below the DAILY Pivot, stop above the DAILY R1, with a MAX
WEEKLY average bearish gain target of $35.86. If QQQ trades
WEEKLY R1, immediately lower stop to protect against gravitation
back to the WEEKLY Pivot.
Jeff Bailey
