The discussion centers on methods for simulating a Chi-squared distribution, particularly when the degrees of freedom (DOF) are very large, potentially up to 10^8. Participants explore algorithms and theoretical implications related to this distribution.
Participants generally agree on the applicability of the Central Limit Theorem to the Chi-squared distribution with large degrees of freedom, but the discussion remains open regarding specific simulation algorithms.
The discussion does not resolve the specifics of simulation methods or the potential limitations of the normal approximation for very large degrees of freedom.
marcusl said:The Central Limit Theorem says that chi-squared approaches a normal distribution when the number of DOF's k becomes large. A normal distribution is an excellent approximation for k>50 in most cases, so it should be near perfect for k=10^8.