values for the MO. The sell-off is preceded by a MO move to negative territory and crossing of the MOS. The MO remains in negative territory throughout the sell-off.

In Figure 4.13, analysis of NASDAQ Composite data shows a slide during the early summer and a partial recovery in late summer. The MO is in negative territory for essentially the entire slide until the double bottom occurs. The MO shows congestion building around the end of July followed by an upward move. This upward movement has stagnated and congestion is once again building. The NASDAQ appears to be preparing for a move.

Steps for Assessing Price mobility

The eight steps for assessing price mobility and trading are described in this section. The numerical methods are derived from the formulas given in the text. "While it is possible to perform the analysis directly on a spreadsheet, it is recommended that macro programs be used to efficiently manage the calculation of the Price Distribution Functions. Visual Basic macro programs with EXCEL were used for the analysis presented here and are described further in the included references.

Figure 4.11 A significant move to the upside in early 1996. There is

evidence of congestion building prior to the large upward move.

Figure 4.12 The rise, sell-off, and bounce of the NASDAQ

composite in summer and fall 1995.

1095

1062

1 29 t

I996

963 i

930 Mobility OsdJUtors

«

ill!!

59 BAR PDF

£

NS 10 BARS M 40 BINS

Figure 4.13 Analysis of NASDAQ composite data. There is a slide

during the early summer and a partial recovery in late summer.

1. Select an issue and a data history of high, low, and closing data.

2. Apply numerical methods that represent the formulas given earlier to compute the mobility oscillator and slow mobility oscillator. Select a look-back period to coincide with your trading horizon. The look-back period should be as long as possible, but short enough to avoid multiple peaks in the PDF. The period should be short enough to be responsive to significant moves (e.g., the value of 14 days for OEX daily data seems to work well). The smoothing period should be shorter than the look-back period, typically one-half to one-third (a value from 5 to 7 days works well for the cited example).

3. Compare high, low, and closing data with MO and MOS on the same time scale. This can be accomplished using EXCEL charts with MO and MOS plotted on the secondary axis. Generally the MO is reasonably smooth; however, sometimes there is an abrupt change. In this case, it is helpful to display the PDF also for more detail as a separate bar chart. Sometimes the abrupt change is meaningful, and sometimes it is the result of multiple peaks in the distribution and abrupt changes in the mode. If there is complex structure, new emerging peaks, or old declining peaks, then be careful. The mode reference point may be switching dis-continuously and an apparent change in momentum may be a false signal.