Monte Carlo Simulations?

Discussion in 'Trading Software' started by monty21, Jun 27, 2009.

Which is better for Monte Carlo simulations?

  1. Matlab?

    3 vote(s)
    33.3%
  2. Mathematica

    2 vote(s)
    22.2%
  3. Other

    4 vote(s)
    44.4%
  1. I'm curious what software I should use for Monte Carlo Simulations. I'm not a computer programmer but have some knowledge of quantitative finance. I assume that Matlab is more sophisticated than Mathematica, but I don't mind investing some time to learn. Whatever is more efficient and better overall will do.

    Please list any other programs that are good as well?

    If anyone has any Excel tips to run a very simply test, that would be great too. I've seen a lot of information on the web for Excel but few for equities.

    Thanks :)
     
  2. One important point to keep in mind is the random number generator. Matlab is good I think but for Excel I dont know.
    Any programming language can do, if you can code it.
     
  3. Thanks for the quick reply... I know there is a random number generator for Excel... but I don't want to deal with too much on the macros to get it to run.

    I would prefer just plugging in data/variables. Does Matlab require a lot of (complex) coding? I was under the assumption that I just insert the data and plug in parameters within a made outline.

    I'll do some more research now.
     
  4. computer can not give you true random numbers

    use a program that measures CPU temp

    it will give you true random numbers
     
  5. Wet dreams.
    You do have to code the thing since you have to define what is it that you do a Monte Carlo on. Matlab does not have a list of all possible testable scenarios.
    And even worse, if you want to test a strategy on stocks, you have to define how the simulated stock path go. You have to make it yourself; brownian? stochastic vol? jumps?...
     
  6. In Monte Carlo you dont care about the numbers themself but rather on their distribution. And it's a desirable feature to have pseudo random number because you can re-run the same simulation many times or run a different simulation with the same numbers.
    But you are right, they are not random, they are pseudo random.
     
  7. timbo

    timbo

  8. I have used @Risk since 1988. At that time it was a Lotus 123 add-in.

    Excellent program!
     
  9. timbo

    timbo

    Wow; didn't know they went that far back.
     
  10. Good post. I wonder if trully random number exist, since they always depend on the seed and thus the sequence will be repeated at some point in time.
     
    #10     Jun 28, 2009