# What free tool can I use to do simple Monte Carlo simulations on OS X?

What free tool can I use to do simple Monte Carlo simulations on OS X?

-
Almost any environment that includes a decent random number generator, I would have thought?! A little more info on the intended application might help elicit a more helpful answer: MC simulations of what? –  onestop Mar 29 '11 at 12:09
You can look at my answers, whenever they contain Monte Carlo simulations, they were done on Mac OS X using R. –  mpiktas Mar 29 '11 at 12:28
CWized, since I thought it will become a dictionary of MC software. –  mbq Mar 29 '11 at 13:29

What is a probability that a sum of a 3 highest results from 5 throws of a dice is divisible by seven?

> mean(replicate(1e5,sum(sort(sample(1:6,5,replace=T))[3:5])%%7==0))
[1] 0.16068
> mean(replicate(1e5,sum(sort(sample(1:6,5,replace=T))[3:5])%%7==0))
[1] 0.16032


Circa 16%.

-

My favourite platforms are

PyMC runs on OS X out of the box, OpenBUGS is originally for windows, but according to this it can be run using Wine.

-
You can use JAGS on OS X (it works with R too). –  chl Mar 29 '11 at 13:15

To some extent (less strictly statistical) NetLogo.

-

You can even do Monte Carlo Simulation in Excel. It's not a perfect tool, but you probably already have it and know how to use it. Depending on the scope of your problem, it might be easier to use Excel than to learn something new. If you are going to learn something new, R is a great choice.

What are you trying to simulate?

-
Keep in mind that spreadsheet software have poor numerical accuracy, see [www.jstatsoft.org/v34/i04/paper](www.jstatsoft.org/v34/i04/paper) . –  GaBorgulya Mar 29 '11 at 18:46
@GaBorgulya This is absolutely true, but if Ionuț needs a quick answer he might be willing to make the tradeoff in accuracy. –  Zach Mar 29 '11 at 19:18