How random is GoldSim?




  • Official comment
    Rick Kossik

    By default, GoldSim repeats the sampling sequence every time you run a model. You can turn off this option (as discussed here in Help:!Modules/5/probabilisticsimulationoptions.htm). Typically, however, you would not turn this off. Otherwise, as you develop (and test) your model, it will be impossible to determine if changes in behavior are due to changes you have made to the model or simply due to random sampling differences. Running a single realization and getting a different answer every time you run it would be confusing and make it difficult to check your logic. Of course, after developing a model that has uncertain inputs, eventually you will need to run multiple realizations. If the number of realizations is sufficiently high, the sampling sequence will not be important (i.e., it won't change the statistics of the outputs). I strongly recommend you read the two Units of the online Course (11 and 12) that explain these things in great detail.  You can read, for example, about the sampling algorithm here: You can read about how many realizations are enough for a particular application here:

    Comment actions Permalink
  • Avatar
    Gioja, Leslie Michael


    Thank you. This was very helpful.

    I have been going through the training am In Lesson 7 now. It was in fact, the training that keyed me into this characteristic of software. It seemed odd to me that , if I ran a simulation with the same parameters as the example and the generation of realizations was random, then I should not get the exact same result for any specific realization (e.g. Realization 1000 in Example 14 in Lesson 7). Your explanation for using the same seed for randomization makes sense and tips on turning it off were helpful.

    Comment actions Permalink

Please sign in to leave a comment.