An underappreciated #rstats simulation option is simple mixtures: eg 80% from N(0,1) and 20% from N(0,5) giving fatter tails "cheaply": > set.seed(42) > X <- c(rnorm(10000,0,1), rnorm(2500,0,5)) > e1071::skewness(X) [1] -0.00247981 > e1071::kurtosis(X) [1] 7.92564 >