The weekly high-low data is taken from ISL (see Chapter Seven), Barrons. the Wall Street Journal, or other available sources. Carefully plotted, the results appear in Figure Vin-2. As usual at this point, little can be said as to whether the stock is a buy," or has more to go on the dovmside, or w4ien.

"Setting Up" Screw And Bolt

17 18 15 14 13 tz

s

SCREW AM>eOtT

1987 (

J J A

FIGURE

Tracing from this master chart we conduct a cyclic state analysis as shown in Figure VIII-3.

A single, constant-width envelope is drawn, and immediately generates pertinent information.

• The price motion up to the last 14 weeks demonstrates magnitude diminuation of a major cyclic component per the principle of variation.

• Two samples of the dominant component are present in the plot - of 20- and 16-week durations respectively. The average for these two samples is 18 weeks. We recognize this duration as consistent with the price-motion model.

• Each such component is seen to be sharply dhrtded into two segments - of seven to 12 weeks in duration (average 9.2 weeks). While somewhat short to compare with the 13.0-week component of the model, we accept the results of our observation subject to verification.

• Each of these is further divided into two segments of four to seven weeks in duration (average 5.2 weeks). This is seen to compare to the 6.5-vreek cycle of the principle of nominality, with the tolerances of the principle of variation.

• The center line of the channel is hard up - telling us that the sum of all components longer in duration than the 18-week cycle is providing up impetus to price-motion.

SEfHESm

SCREW AND BOLT FK3UREia[-3

A 4*4

SERIES ---12

SERKSI

-20-

INVERSE 10 WK. MOVING AVERAGE

4 -

SERCSn SERIES I

The Weekly Prediction Plot

Because of the rather large variation in observed duration from model nominal, we check our conduaons using an inverse moving average. A span of 19 weeks is chosen (18 would be ideal, but being an even number would require inte olation before subtraction from corresponding prices). The 19-week moving average of the weekly mean prices is formed and subtracted from the corresponding means. You will recall that the effect of this operation is to display about a horizontal base line the sum of all components shorter in duration than 19 weeks. The result is shown at the bottom of Figure VIII-3. With minor differences, the results of observation are found to be confirmed.

The next step is to extrapolate the average (and variation from average) durations of the three periodicities found, identified as Series I, , and III in Figure VIII-3. To do this, we measure forward in time from the last identified low of each, and obtain the overlapping shaded areas marked I, II, and III. This is an expected nest of lows, and is seen to extend from present time to about three weeks into the future. We expect a significant bu3ring opportunity in this stock-soon!

Having passed the initial analytical tests as an issue of interest, we now wish to construct our "working" chart. Gathering daily data from the last low of the 18-week cycle and plotting produces Figure VIIM. The envelope and nest of lows are reproduced on this chart from the preceding figures.

As expected, going to daily data produces new information regarding still shorter duration components. The one that is most dominant is eight to 11 days long, or an average of 9.9 days. We are now ready to state the cyclic condition of the stock for the following week as follows:

SCAEW AND

neuReva-4

I ll

-VTt-l

The DaOy Predkjtion Plot