PureBytes Links
Trading Reference Links
|
Philip!
When you calculate your daily differences you are most likely using only
Working Days, probably between 250 and 260 observations per annum. Hence
try employing the square root of time rule using the actual number of
observations per year, e.g.
Std(Log(C/Ref(C,-1)), X)*Sqrt(250)*100
Also you can check whether there is a constant ratio between your
calculated volatility and the volatility of your other source. If this is
the case then the above is probably the explanation.
Also there are other measures of historic volatility. E.g. the measure of
volatility may be "zero based", it may be employing exponential moving
averages instead of SMA, etc.
Best of luck,
Yngvi
At 12:32 PM 26-10-98 -0500, you wrote:
>I have compared historic volatility as calculated
>by the Metastock formula
>
>Std(Log(C/Ref(C,-1)), X)*Sqrt(365)*100
>
>with the values published at another source and
>found disparities. Can anyone offer an
>explanation? Examples follow:
>
>Differences in the March beans are pronounced:
>
>MS hisvol(7) = 10.46
>PMoptions page “1-week” hisvol = 8.44
>
>MS hisvol(90) = 26.76
>PMoptionspage “3-month” hisvol = 16.52
>
>Differences in December CD are less striking:
>
>MS hisvol(7) = 4.02
>PMoptions page “1-week” hisvol = 4.26
>
>MS hisvol(90) = 10.78
>PMoptions page “3-month” hisvol = 11.17
>
>An email inquiry to the PMoptions page has not
>been answered.
>
>. . . while we’re at it, does anybody know of
>other free sources for hisvol on the web?
>
>Philip
>
>
>
|