by Ken Long
One of the most powerful measurements in the technical trader’s toolbox is Average True Range, invented by Welles Wilder, and found prominently near the top of every major charting package that can be found on the web. Its power is what makes it so popular.
I always recommend to traders that they use ATR for their systems if they need a decent measure of short term volatility that is sensitive enough to detect important changes in near term volatility, smooth enough to prevent whipsaws of opinion and action, and reliable enough to be useful in all types of market conditions.
The way the indicator is calculated and displayed though makes it very easy to misinterpret by a trader or investor who is not familiar with how it works. In my opinion, the problem is so profound that it should not be displayed in the raw form on your charts because the visual display of the information is so misleading. Even experienced traders, who look at a lot of indicators, can overlook the problem. In fact the more indicators you use, the easier it is to forget this insight about ATR.
The problem is simply that ATR is measured in dollars and cents and so when you see the ATR line rising on a chart you are naturally led to believe volatility is increasing. The problem is that if price is also rising, volatility may be steady or even declining, which you cannot easily tell by inspection.
For example: an ETF priced at 20 with an ATR of 2 is much more volatile than an ETF priced at 100 with an ATR of 5. If it were the same ETF, and you watched its ATR climbing from 2 to 5 while price went from 20 to 100 you might assume that it was more volatile as price was climbing. The opposite would be the truth of course, but your eyes could lead you to believe otherwise unless you were constantly alert.
Here is the refinement that I recommend. Instead of using ATR in its basic form, simply divide ATR by the current price. This will give you a number which I call “ATR%”. This is interpreted to describe the percent fluctuation of the value of the asset that can be considered normal given recent price action. By recent price action I mean the last 14 trading days, in the default construction of the indicator.
Use of the ATR% will allow you to do the following things very easily, things which cannot be done using straight ATR numbers, which represent a dollar denominated “normal” fluctuation.
You can use a time series of a single asset to see how its volatility is truly changing over time.
You can compare the ATR% of an asset against a set of other assets, and rank them into a scheme of relative volatility at that moment.
You can use statistics to further classify and describe a set of assets into logical, statistical sets.
You can compare sets of volatilities against other time periods to help you with market classification.
Average True Range is an important tool of technical analysis, but one which has serious problems in interpretation for novice or very busy traders. Know your tools.