This function returns the Garman-Klass estimation of volatility. It was developed by Graman and Klass and it uses the high, low and close prices to estimate volatility.
The Garman-Klass estimator is up to eight times more efficient that the close-to-close estimator and unlike the Standard deviation and the Parkinson estimator, it uses three price information, which are the close, high and low prices.
It also accounts for the opening jumps or gaps in the price series by using the close and previous close prices.
However, it is more biased that the Parkinson estimator.
(From 'Volatility Trading' book)
The close-to-close volatility estimator is simply the standard deviation (Square root of the variance) of returns (QuantShare function: Stddev).
Another measure is the Parkinson estimator, which do not use the close price in its calculation. Instead, it uses the high and low prices. (The function that returns this volatility measure can be downloaded here: Historical High-Low Volatility: Parkinson).
NB: The function accepts only one parameter, which is the number of bars to use to estimate volatility.
Annualized Garman-Klass volatility is calculated by multiplying the return value by the square root of the number of trading periods in a year.
