[A new charting technique: Volatility Limit Chart]
Although there are many methods of defining volatility, for short-term time frames (e.g. 7 days, 5 days, etc.), it is generally accepted that the use of standard deviation of closing prices is not valid for these cases due to the small number of samples in the observation window. The daily range is often used as a proxy for volatility (some believe it to be a better indicator of volatility than the std-dev of closing prices), which leads us to the volatility limit chart (VLC) concept (which includes open-high-low and close data). By plotting the highest high and lowest low over an observation window, and using the opening price on the first day of the window, and the closing price on the last day of the window, then moving the window 1 day forward and repeating this process ad-nausium across the chart, we have a new chart technique that I have been experimenting with for a while. Since I had been experimenting with the charts in candlestick format, and since Nison refers to the process described above as a blended candle in his book (except he applies it in isolation to a single group of 3-5 consecutive candles), I was calling this a Moving Blended Candle chart. The volatility angle hadn't occurred to me until recently (probably because market volatility is much higher of-late, compared to several months ago when I first started looking at this-so it jumps out more now). Essentially, the VLC shows, on any given day, a running tally of the n-day composite volatility, as measured by the highest high and the lowest low price in each window position. Since it is a running tally of the "n" latest trading days, range breakouts are easily seen due to it's inherent "peak hold" operation. It's also very good at pointing out opportunities for swing trades (by expected time and price) that aren't always apparent from a regular chart. The open-to-close range on any given day shows the net price progress over the trailing n-days. If, for example, the body of the candle is small, but the upper and lower shadows are very long, the stock's making a lot of noise with no real price progress. Because it is blending the data by pulling the trailing n-days across the chart and keeping the extremes, it sort of "smears" the data from right to left. But, unlike a linear filter (e.g. moving average), there is zero lag in the volatility chart. All data in the volatility chart occurred exactly as shown at some point in the chart (not scaled in magnitude like a moving average). I'm finding Wilder's Average True Range is an interesting indicator to apply to the VLC, as is a custom accumulation/distribution of the n-day total volume (since that is the total volume that drove the volatility) . Unfortunately, backtesting it would take a lifetime because I have to implement this in excel, and transfer it back for each stock I look at. I think a person who uses candlestick charts would get more value out of it than a bar chart'er. If anyone is interested in testing it out, I can send you the Excel file.
dh |
