This author states that, while the article by Gunter Maris and Timo Bechger ("On Interpreting the Model Parameters for the Three Parameter Logistic Model," this issue) is highly interesting, the interest is not so much in the practical implications, but rather in the issue of the meaning and role of statistical models in psychometrics and educational measurement. Maris and Bechger show that the same model in two parameterizations (the parameterization of George and the parameterization of Frederick) could lead to opposite conclusions with respect to whether students guess or not. In this article, Glas discusses two points: (1) that the problem raised by Maris and Bechger is omnipresent in psychometric modeling for psychological purposes and cannot be resolved given the models and the data that have generally been available at the moment; and (2) that this does not bother the author because the models are highly useful in educational measurement for very different purposes other than for psychometric modeling for psychological purposes. In conclusion, Glas comments that reading the article by Gunter Maris and Timo Bechger has been very valuable, because it served as a reminder of what can and cannot be done with IRT.