Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.


Paul
Posts:
769
Registered:
7/12/10


Reporting results of Monte Carlo simulations
Posted:
Jun 14, 2014 2:30 PM


Suppose that a research paper analyses probability distributions via Monte Carlo simulations, and then reports on the results. Is a reputable research paper supposed to report the findings in such a way as to ensure that other authors can replicate it? For example, instead of just saying that a sample of 100000 points were taken from an N(0,1) distribution, perhaps the paper should say exactly what IDE and computer program was used to generate the points and which random number seed was used.
The reason I ask is that I read a Monte Carlo paper which described such a simulation but gave no help to readers who wanted to replicate the results  no code was included, no indication of random number seeds etc. I tried and failed to replicate the results.
Is the omission of complete detail about exactly how the simulations are conducted standard practice?
Thank You,
Paul Epstein



