Simulations

[nick.smith99]

I know how to simulate normal or lognormal returns within VBA using the RND function, and normsdist within the worksheetfunction bit of VBA (there may be more efficient ways but I know this works), but does anyone know how to incorporate kurtosis within the code. In other words can I simulate a distribution with a known kurtosis number using excel/vba?

 

[shg]

You could generate deviates from a distribution that has a kurtosis you like.

 

[nick.smith99]

Thanks for the response. Just so I understand, do you mean take a dump of data that has some kurtosis and than add some "noise" to it to create a distribution with kurtosis. That makes sense, although I was trying to play with the kurtosis number to see what effects that has.

 

[shg]

I'm saying you could generate variates with, for example, a logistic distribution, or a hyperbolic secant distribution, both of which have kurtosis greater than a normal distribution, but look more or less normal-is.

 

[nick.smith99]

Thanks. I will try those out.