Using GRE quant score and GPA score as proxies for student selectivity is a bit of problematic. For the top programs, their students' GRE quant scores are all clustered slightly below 800. In regression models, we know that a variable with very small deviation tends to serve as "bad" explantary variable since it would introduce huge standard error for estimated coefficient. Similar reasons can be applied here to argue that the usage of GRE quant score would introduce huge measurement error as a proxy for student selectivity.
For GPA score, we need to keep in mind that it is much more difficult to get 3.5 GPA from an Ivy school than from a third tier university. It is almost certain that better MFE programs tend to accept students from better schools. It is not unthinkable that MFE students from better school may actually have lower GPA on average than students from not-so-good schools because of this. Therefore, using average GPA without controlling for the type of universities from which students had their bachelor degrees is probabily not a good way to measure student selectivity either.
Last but not least, I'm not sure student selectivity deserves a 25% of weighting for the ranking even if good proxies for it can be found. Once the performance of students after graduation is controlled (as is in the case of this ranking), one would think that student selectivity is negatively correlated with the quality of the program. A way to think of this is to imagine that there are two programs-- one accepts all the crappy students and one admits all the talented students. But after graduation, the students from both programs manage to earn the same income and enjoy the same reputation in the industry over the years. Now which program do you think is better, the one that gets all the crappy students or the one that gets the talented ones? From the view point of employers, this doesn't matter. However, from the point of a student, the first program is actually the better one.
That being said, I don't mean to bash this ranking. I'm just trying to give some constructive criticism here
Personally, I like this website as it has provided me with many useful information. I also respect Baruch. I attended ARPM bootcamp held in Baruch and was very much impressed by what I saw there.