the hAndbook of technIcAl AnAlysIsbandwidth
The bandwidth, that is, the distance between the upper and lower bands, may be
based on any of the following:
■ (^) Central value ± percentage of the central value
■ (^) Central value ± fixed dollar or point value
■ (^) Central value ± multiples of standard deviation based on the central value
■ (^) Central value ± multiples of ATR based on the central value
Here are some popular examples of bands constructed based on measures of
volatility and percentage:
■ (^) Bollinger Bands: 20‐period SMA of closing prices ± 2 standard deviations of
the 20‐period SMA
■ (^) Fixed Percentage Bands: Any moving average ± a fixed percentage of the cur-
rent moving average value
■ (^) Fixed Value Bands: Any moving average ± a fixed dollar or point value of the
current moving average value
■ (^) Keltner Bands: 10‐period SMA of typical price (H+L+C)/3 ± 10‐period SMA
of ATR; some modern variations include the use of a 20‐day SMA or even an
exponential moving average (EMA) as its central value
■ (^) STARC Bands: 6‐period SMA of closing prices ± multiples of a 15‐period
ATR; multiples are normally two or three times the 15‐period ATR
■ (^) ATR Bands: Any moving average ± multiples ATRs of its central value
band sensitivity
An envelope or band may be adjusted for greater or lesser sensitivity with respect
to price action by:
■ (^) Adjusting its periodicity (i.e., lookback period)
■ (^) Selecting the type of averaging to employ for its central value
■ (^) Adjusting the bandwidth or distance of the bands from its central value
Figure 12.5 shows the sensitivity or responsiveness of various band periodici-
ties and bandwidth percentages to price action, on the daily chart of the Technol-
ogy Select Sector SPDR ETF.
Figure 12.6 shows the sensitivity or responsiveness of various bands to price
action, with different averaging methods.
bandwidth bias
The size of the bandwidth may or may not vary with price, depending on the type
of containment employed. The only containment method that guarantees an in-
crease or decrease in the bandwidth as prices rise and fall is the fixed percentage
band. There is a bias toward larger bandwidths at higher prices because the band-
widths are based on a percentage of the central value that is greater at higher