**The Keltner Channel and Bollinger Bands**

(Please also find an AFL for both Keltner Channel and Bollinger Bands in AFL Section of the forum.)

Channels and bands of various origins have been used to study market price movement by day traders from many disciplines.

The best known day trader who utilizes the Keltner Channels and has published some articles on the topic is Linda Bradford Raschke. Without quoting her verbatim, if you have the multiples set up for a particular day and most, say 90% of the price action stays within the channel, you would be able to spot overbought and oversold signals to work around. But this explanation also points out what is, for me, the real weakness in using Keltner Channels.

How do you know, on a daily basis, which multiples of the moving average to use and, for that matter, what time frame is appropriate for the moving average itself. I suppose with years of experience you might develop the ability to judge the market and set the appropriate variables, but it sounds like a fairly tall order for a novice trader. Raschke has done work integrating the Average True Range indicator into the moving average with some success, which seems a more accurate methodology to my way of thinking. The point is simple, though; the Keltner Channel methodology would take some very specific mentoring to be an effective trading tool for your indicator set. At best, it serves as a nice filtering device for other primary trading indicators.

There are three steps to calculating Keltner Channels. First, select the length for the exponential moving average. Second, choose the time periods for the Average True Range (ATR). Third, choose the multiplier for the Average True Range.

The first step in calculating the Bollinger Bands is to find the simple moving average. The upper and lower Bollinger Bands are calculated by determining a simple moving average, and then adding/subtracting a specified number of standard deviations from the simple moving average to calculate the upper and lower bands.

There are two differences between Keltner Channels and Bollinger Bands.

First, Keltner Channels are smoother than Bollinger Bands because the width of the Bollinger Bands is based on the standard deviation, which is more volatile than the Average True Range (ATR). Many consider this a plus because it creates a more constant width. This makes Keltner Channels well suited for trend following and trend identification.

Second, Keltner Channels also use an exponential moving average, which is more sensitive than the simple moving average used in Bollinger Bands.

Instead of using a preset multiple of the SMA the Bollinger Bands set the outer lines at two standard deviations from the center line. The level of standard deviation can be altered, but the generally accepted norm seems to be about two standard deviations. So we are dealing with a non-linear outer line formation now, since the standard deviation changes in size depending upon the position of the center line. When the market is consolidating, Bollinger Bands tend to draw very close together, showing a very low level volatility. Conversely, when the volatility is increasing, the bands will swing wildly away from each other and the width between the outer bands becomes greater.

Contrasting this with the more equidistant demeanor of the Keltner Channel will immediately show the casual day trader the difference in these two indicators. One is linear, one is non linear, and the reality is that they appear very different on a chart.

Oddly enough, though, I consider the Bollinger Band to be a secondary indicator, though there are trading systems that use them as a primary indicator. The general rule of thought on both the Keltner Channel and Bollinger Bands is fairly simple: a close outside the channel is indicative of overbought and oversold conditions and hence, there is a potential counter-trend trade in the offing.

My experience has been that the Bollinger Bands are more accurate at predicting countertrend moves. Again, I would use them as a filter device and see if, in fact, my primary indicators show the same information. I think you could say the same for the Keltner Channel, which I have used less.