.BG
.FN Stable
.FN rstab
.TL
Stable Family of Distributions
.CS
rstab(n, index, skewness=0)
.AG n
sample size.
If `length(n)' is larger than 1, then `length(n)' random values are returned.
.AG index
vector of indexes in the interval (0,2].
These are specified in the form given
in the reference.
An index of 2 corresponds to the
normal, 1 to the Cauchy.  Generally, smaller values mean
longer tails.
.AG skewness
vector of modified skewnesses (see the reference).  Negative
and positive values correspond to skewness to the left and right.
.RT
random sample from the specified stable distribution.
.SP
.PP
Stable distributions are of considerable mathematical interest.
Statistically, they are used mostly when an example of a very long-tailed
distribution is required.  For small values of `index', the
distribution degenerates to point
mass at 0.  See the reference and other works cited
there.
.PP
Note that there are no density, probability or quantile functions
supplied for
this distribution. The efficient computation of such values is
an open problem.
.SH REFERENCE
J. M. Chambers, C. L. Mallows, and B. W. Stuck,
"A Method for Simulating Stable Random Variables",
.ul
JASA,
Vol. 71, pp. 340-344, 1976.
.EX
hist(rstab(200,1.5,1.5)) #fairly long tails, skewed right
.KW distribution
.WR
