By Linda Piazza

Bollinger bands, envelopes, channels -- whatever your preference -- catching moves from one boundary to another can be profitable. What about those times when prices slide along a band or envelope instead of rebounding toward the opposite side? No one wants to be stuck on the wrong side of a trade while prices slide along a slippery slope.

OEX 30-Minute Chart with 1.35% Envelope Surrounding a 21-pma

Is there a way to identify those times when prices will likely slide rather than rebound? What affects the rate at which prices rebound? One possibility appears obvious: The slope of the envelope, channel or band at the time price hits it.

Test 1: The Effect of the Slope of the Envelope, Channel or Band on When Prices Will Slide
To test that thesis, I used a 1.35% envelope surrounding a 30-minute 21-pma on the OEX. I gathered data for each first touch of that outer envelope for a year, beginning in October 2002 and continuing through the end of September 2003. I calculated slope as a change in envelope value (y) divided by a change in time (x).

The y-portion of the slope was the difference between the value of the touched envelope one bar before the touch and at the time of the touch. I divided that result by the time span. That established a slope with a price-per-minute label.

Test 2: Determining the Strength of the Rebound
Determining the strength of the rebound proved more difficult, but, for this test, I used the number of 30-minute bars between a first touch of the outer envelope and the next touch of the central 21-pma as the most objective and easily calculated measure.

Interestingly, the year’s data resulted in an even number of data collection points for the upper and lower envelope boundaries: 48 for each. The average number of 30-minute bars between the first outer envelope touch and the next central average touch calculated to be 11.83 bars, almost six hours.

Separating touches of the upper envelope boundary from the lower boundary demonstrated some differences. Touches of the lower envelope boundary resulted in an average rebound of 10.79 bars’ duration. Touches of the upper envelope boundary resulted in an average rebound of 12.88 bars’ duration.

While it may prove helpful to know how many 30-minute bars it might usually take the OEX to rebound to the central average, that wasn’t the subject of this inquiry. This test hoped to pinpoint the impact of the envelope’s slope on the quickness of the rebound.

If I’d had any expectations previous to the test, I would have expected prices to rebound most sharply when slope approached zero (the envelope band was horizontal) and prices to slide if the channel sloped at the time of the touch. I would have expected the data to show a critical point at which the slope would have been too steep, and a rebound would also tend to occur.

Using a scatter chart to display all 96 data points revealed a different-from-expected concentration. If my original assumptions had been correct, the scatter chart would have been shaped like an elongated eye mask, with data points pinched in as they concentrated at a low number of rebound bars (quick rebound) for a slope near zero, widening on either side of zero, and then pinching in again as slope moved farther away from zero in either direction.

Instead, the shape was vaguely like that of a megaphone, concentrating near 10 when slope was most negative and widening as slope grew most positive.

Scatter Chart Showing All Data Points

An analysis of this chart reveals that when the envelope sloped down (negative slope), rebounds gathered fairly reliably near 10. When the envelope sloped up (positive slope), rebounds scattered more widely. I could find no correlation between slope and the strength of the rebound, at least when the slope was positive. No concentration of data points pinched together near zero, depicting quick rebounds.

Because rebounds concentrated differently when slope registered negative values than when it registered positive values, the next question to be asked was whether rebounds concentrated differently after hitting the upper and lower envelope boundaries.

A scatter chart of data points collected during the 48 touches of the lower envelope demonstrated some of the same patterns seen in the chart of all data points. When slope registered negative values, data points tended to collect near the calculated 10.79 average for touches of the lower envelope boundary. When slope registered positive values, data points scattered more widely.

Scatter Chart Showing Touches of the Lower Envelope

In contrast to the pattern seen on the previous chart, the scatter chart depicting the 48 touches of the upper envelope showed no discernable pattern. These data points scattered widely with no particular concentration

Scatter Chart Showing Touches of the Upper Envelope

After examining the charts depicting lower and upper envelope touches, I concluded that, when the envelope slanted down (negative slope) and prices touched the lower boundary, rebounds tended to be faster.

Test 3: The Effect of Bullishness on the Speed of Rebounds
Thinking about that tendency in relationship to the bullishness that characterized some periods of this year, I wondered if the quick rebounds off the lower envelope boundary owed more to that bullishness than to anything relating to the slope of the envelope. To test that thesis, I divided the year’s data points into four quarters: Q4 2002, Q1 2003, Q2 2003 and Q3 2003.

Q4 2002 showed a heavy concentration of data points near and below 10, showing relatively quick rebounds from both upper and lower boundaries.

Scatter Chart Showing Q4 2002

A volatile quarter, Q4 2002 produced numerous touches of both the upper and lower envelope boundaries, with prices dropping quickly early in the quarter and then climbing steeply after that drop.

OEX Daily Chart of Q4 2002

Because the values on the Q4 2002 scatter chart proved similar as slope varied from -2 to +2, the quick rebounds did not seem dependent on the slope or on whether the upper or lower envelope was touched but, rather, seemed a product of the volatility of the markets.

Performance during other quarters backed up this observation. For example, many traders commented on the range-bound trading in Q3 2003. This quarter produced fewer touches of the envelopes, so fewer data points. This quarter also produced the most widely scattered data points.

Scatter Chart of Q3 2003

Although the various scatter charts of all data points across the entire year showed some correlation between quick rebounds and touches of the lower envelope boundary when the envelope sloped down, that result appeared to be due to Q4 2002’s results, varying widely from quarter to quarter.

The Good News and the Bad
The inevitable conclusion was an obvious one: When markets are volatile, they rebound from outer envelope boundaries more quickly. Unfortunately, instead of slope-dependent results, these tests revealed volatility-dependent results. It may, therefore, be dangerous to draw too many conclusions about likely price action based on whether an envelope, Bollinger band or channel slopes up or down, a conclusion many of us are prone to draw.

That’s the bad news. The good news? We don’t have to spend a whole lot of time diligently calculating slopes during fast-moving markets.

*Reprinted (and modified) with permission from OptionInvestor.com