The problem is, ultimately, university acceptance is one dimensional. Fundamentally, the university is going to analyze all its applicants, more-or-less rank them according to its standards, and then take the top X applicants according to its metrics. Since you can't half-accept someone, there isn't a way to do anything else.
And once you have that measure, you are now subject to Goodhart's law [1]. There is no escape.
So, your university gives points for applicants who have extracurricular activities? Here comes someone that joined 15 clubs. Oh, do you want them to be leaders in their activity? Here comes someone who is the president of six clubs, and founded two more. Oh, do you want them academically accomplished instead? Here comes someone who in high school published six scientific papers (please don't look too hard at which journals they were in). No, wait, we don't like what that's encouraging, let's just look at standardized test scores... and here come the perfect scores.
Applicants will make themselves one dimensional. They'll tune to whatever dimension you're measuring on, no matter what you try to do to the basis vector of that dimension. And they'll beat out anyone who is just being normal, or even more cynically, just being honest about their actual activities.
One of my favourite suggestions to counter this from an essay on overfitting I really like [1] is that applicant's probability of being accepted is proportional to their position in the ranking.
The downside is that some great applicants will not enter their top choice of school, and some people who aren't great fits will. But, on the other hand, the perverse incentive of spending the entirety of high-school optimising for some arbitrary metrics will dissappear. Any marginal improvement is corralated with a marginal increase in success, rather than the current system of no pay-off whatsoever, until reaching some arbitrary threshold, where one gets all the pay-off at once.
> The problem is, ultimately, university acceptance is one dimensional.
Ultimately all things are reducible to a function applied to the state of the world. The entire point is that by introducing additional dimensions to the metrics of 'acceptable for admission' the 'representation of the student' ("state of the world") is made out of a higher dimensional variable.
Now, you can game the "one dimensional" student representation or the n-dimensional version. For each of the possible (predictable, finite set) attacks on the function there can be remedies. where there are cases where we can not have effective remedies, that fact itself appears to be independent of the number of dimensions.
Ultimately, it seems hard to argue against a higher dimensional metric given that it is more information rich. And, I venture to guess young teens would also welcome the mere possibility of 'choices' in electing areas to excel.
And once you have that measure, you are now subject to Goodhart's law [1]. There is no escape.
So, your university gives points for applicants who have extracurricular activities? Here comes someone that joined 15 clubs. Oh, do you want them to be leaders in their activity? Here comes someone who is the president of six clubs, and founded two more. Oh, do you want them academically accomplished instead? Here comes someone who in high school published six scientific papers (please don't look too hard at which journals they were in). No, wait, we don't like what that's encouraging, let's just look at standardized test scores... and here come the perfect scores.
Applicants will make themselves one dimensional. They'll tune to whatever dimension you're measuring on, no matter what you try to do to the basis vector of that dimension. And they'll beat out anyone who is just being normal, or even more cynically, just being honest about their actual activities.
[1]: https://en.wikipedia.org/wiki/Goodhart%27s_law