I have tried different distributions on my data and found out that inverse gamma distribution fits the best. I want to perform a goodness of fit test to get a significance value. I wonder which goodness of fit test (Anderson darling test, Cramer-von Mises or Kolmogorov-Smirnov) is appropriate for inverse gamma distribution. I have tried gofstat function in R but it mostly returns not computed for Anderson darling test and I am not sure what is the reason behind it.