I'd like to hear your opinions concerning which parameters you'd report when estimating some different likelihood based regression? AIC, BIC, Pseudo R^2? What is standard to report? It should be a parameter which gives the answer on, how good the model is specified.

Regards