what is the criterion for rejecting or accepting the fitted distribution?

[qtbgo]

[ FlexSim 24.2.0 ]

In FlexSim 24, the EmpiricalDistribution tool can be used to fit data and determine the statistical distribution function of the data. However, what is the criterion for rejecting or accepting the fitted distribution? It appears that FlexSim does not provide this criterion. Previously, when using ExpertFit, it did offer a standard for accepting or rejecting the fitted distribution.

[support5CRPZ]

You can use 'Prediction error (w)' value.

[qtbgo]

How do you use the Prediction error (w) value as a criterion for judgment? Is there a commonly accepted threshold such that when the Prediction error exceeds this threshold, the fitted distribution is rejected? If so, what is this threshold value?

I hope to obtain the p-value.