To keep this on topic with HN, how could engineering / math / software could inform the admissions process? In my opinion, ranked lists in general are misinterp

To keep this on topic with HN, how could engineering / math / software could inform the admissions process? In my opinion, ranked lists in general are misinterpreted by nearly everyone. Here's an example. Suppose I create an admissions exam and admit the first N top-scoring students. Great, that might explain a great deal of admissions criteria among many schools, seems reasonable. Now, administer the exam by putting the hardest questions first (i.e. lowest correct answer rate in some representative sample). Enforce questions must be answered in order, and immediately terminate the exam after a student has gotten some number K of answers wrong. Admit the students who answered the most right questions. What would the admit list look like? It would turn out to look basically the same as the conventional way the test was administered, except with some noise near the cut-off number of questions--in other words, K and N are the same parameter in disguise. What would be different? For something like the Specialized High School Admissions Test or how the SAT is used for highly competitive schools, the vast majority of students would get every one of the first K questions wrong and be immediately eliminated. You can't possibly claim that their knowledge or competency was tested, because so few questions can't possibly be representative of comprehensive knowledge or competency. And yet, the admit list would be nearly the same! I can say, with high confidence, that the purpose of at least test-based admissions is not to validate knowledge or competency. That in fact, making students endure taking the whole test is a waste of their time. And now it should be obvious why such an admissions structure, despite being what some people and some schools want, feels so immensely unfair. And the only way I came to this conclusion is by just looking at the math.