|
PureBytes Links
Trading Reference Links
|
Assuming you are doing a Monte Carlo dispersion analysis, I would suggest
using actual price dispersions that you randomly sample. This would give a
different answer from the usual option pricing models that assume a Gaussian
distribution rather than a "fat tail" chaotic distribution.
I assume that you are doing an analysis to discover pricing inefficiencies.
I guess I should look up the Barrons article.
Marlowe
----- Original Message -----
From: M. Simms <prosys@xxxxxxxxxxxxxxxx>
To: Marlowe Cassetti <marlowec@xxxxxxx>; Omega-List <omega-list@xxxxxxxxxx>
Sent: Thursday, January 20, 2000 10:11 AM
Subject: RE: Concept & calculation for "dispersion"
> Thanks for the feedback.
> This sounds like a binomial distribution or a skewed dist curve or
> something.....(my stats are foggy !).
>
> Do I need to perform curve-fitting on the actual price distribution or do
I
> use the price deltas or
> better yet, the price-relatives in percentage terms ?
> There must be a statistical measure for this "skewness"....no ?
>
> The issue is how to absolutely MEASURE dispersion....that is the trick.
> Then it can be compared to volatility or implied volatility.
> The ratio of the two is then an important indicator.
> > -----Original Message-----
> > From: Marlowe Cassetti [mailto:marlowec@xxxxxxx]
> > Sent: Wednesday, January 19, 2000 12:50 AM
> > To: M. Simms; Omega-List
> > Subject: Re: Concept & calculation for "dispersion"
> >
> >
> > let me take a crack at this since I was involved in "dispersion
analysis"
> > during the Apollo Program.
> >
> > Assume you construct a Monte Carlo dispersion analysis or simulation of
> > options pricing and you crank it for 10,000 trials. As inputs
> > you randomly
> > sample past price data. Your results could be plotted in a frequency
> > histogram which could be very different from the standard bell shaped
> > distribution. This type could give you an insight into a better
> > description of options pricing based on past price dispersion of the
> > underlying security.
> >
> > Just a guess ... Hope it helps ... Marlowe
> >
> >
> > ----- Original Message -----
> > From: M. Simms <prosys@xxxxxxxxxxxxxxxx>
> > To: Omega-List <omega-list@xxxxxxxxxx>
> > Sent: Monday, January 17, 2000 11:41 AM
> > Subject: Concept & calculation for "dispersion"
> >
> >
> > > For all of you options people.....
> > > the recent "Striking Price" article in Barrons mentioned the concept
of
> > > dispersion which is
> > > supposed to be different than volatility.
> > >
> > > A guy named Carpenter indicates that dispersion has prediction
> > power....when
> > > used along with VIX.
> > >
> > > Has anyone seen this in practice ?
> > >
> > > How can dispersion be calculated ? It seems to be a sort of
directional
> > > volatility....
> > > if all components are girating in the same direction, then dispersion
=
> > > zero, but volatility could still be high.
> > > Dispersion would be greatest when half the components are going up and
> > half
> > > are going down....
> > >
> > > interesting.
> > >
> >
>
>
|