|
Garman-Klass Volatility
Issued on September 11 2011 par Strategies-options.com
First Close to close, then Highs and lows. Next both with opening price added !
First we have seen the close to close volatility ( Volatility : Trading Formulae ) . We then have seen Parkinson's one ( Parkinson Volatility ). The obvious next step is to link both of them. With another data.
I - A simple fact to take into account
The previous formulae are correct as far as the market is a continuous one. Forex seems to be the right candidate.
But even with forex trading, there are some gaps. It's straightforward to take into account this feature by adding another data : the opening price.
That's Garman-Klass Volatility.
II - A formula, a better efficiency
The next formula is 7.4 more efficient than close to close volatility, that is, 7.4 less data are needed to reach the same accuracy.
The formula :
For n days, Garman-Klass variance is
σ²(GK) = ( 1/n ) . ( ∑ [0.511 . ( ln( Hi/Li) )² - ( 0.019 . ln( Ci/Oi ) . ln( HiLi/(Oi²) ) ) – (2 .ln( Hi/Oi ) . ln( Li/Oi ) ) ] )
With,
Oi opening price for the ith day
Ci closing price for the ith day
Hi Highest price for the ith day
Li Lowest price for the ith day
Ln is the natural logarithm
Hence, Garman-Klass Volatility is :
σ (GK) = √ [ 252 . ( 1/n) . ( ∑ [0.511 . ( ln (Hi/Li))² - ( 0.019 . ln(Ci/Oi) . ln(HiLi/(Oi²))) – (2 .ln(Hi/Oi) . ln (Li/Oi))] ) ]
If there are 252 trading days to annualize that number.
III - Example
Using Open, High, Low and Close prices on CAC 40 Index, it leads :

Next : Implied Volatility
Previous : Parkinson Volatility
Related Articles :
- Garman Klass Volatility link : On the Estimation of Security Price Volatility from Historical Data
- The Garman-Klass volatility estimator revisited
- Modelling Volatility Using High, Low, Open and
Closing Prices: Evidence from Four S&P Indices
OPTIONS 101 - INDEX
OPTIONS 101 - CHAPTER I
OPTIONS 101 - CHAPTER II
OPTIONS 101 - CHAPTER III
OPTIONS 101 - CHAPTER IV
OPTIONS 101 - CHAPTER V
OPTIONS 101 - CHAPTER VI Strategies-options.com
|