Kaufman Adaptive Bands

2019년 8월 6일

Introduction

Bands are quite efficient in technical analysis, they can provide support and resistance levels, provide breakouts points, trailing stop loss/take profits positions and can show the current market volatility to the user. Most of the time bands are made from a central tendency estimator like a moving average plus/minus a volatility indicator. Therefore bands can be made out of pretty much everything thus allowing for any kind of flavors.

So i propose a band indicator made from a Kaufman adaptive moving average using an estimate of the standard deviation.

Construction

The Kaufman moving average is an exponential averager using the efficiency ratio as smoothing variable, length control the period of kama and in order to provide more smoothness a power parameter has been introduced, higher values of power will return smoother results.

The volatility indicator is made from a biased estimation of the standard deviation by using the square root of the mean of the square minus the square of the mean method, except that we use kama instead of a mean.

The bands are made by adding/subtracting this volatility indicator with kama.

How To Use

The ability of the indicator to adapt to the current market state is what makes him a great tool for avoiding major exposition during ranging market, therefore the indicator will have a greater motion during trending market, or more simply the bands will move during trending markets while staying "flat" during ranging ones. Therefore the indicator might be more suited to breakouts, even if some cases will return what where turning points, this is particularly true during ranging markets.

Of course the efficiency ratio is not an "unbiased" trend metric indicator, it can consider high volatility markets as trending markets. Its one of his downsides.



High values of power will create smoother bands.



When using a low power parameter use an higher mult. In general using a low power value will make the bands move more freely as well as making them closer to each others.

Conclusion

At least the indicator is really nice to the eyes when using high power values, its ability to adapt to the market is a great addition to other more classical bands indicators, i also introduced a volatility estimator based on kama, some might have used the following estimation : kama(abs(price - kama)) which would have created a slower result. A trailing stop might be made from it if i see request about such addition.

If you are curious here are some more images of the indicator performing on different markets. Thanks for reading !