I definitely think when looking at just RD admissions to selective (and not just the most selective) colleges in the most popular markets, it has become systematically less predictable what those colleges will do with a kid with a given set of numbers, including what may strike people as a lot of āanomalousā results where kids will get waitlisted or rejected from a supposedly less selective colleges while being admitted to supposedly more selective colleges. I personally think this makes sense once you understand the complexities of holistic review, the possibility of different institutional priorities, different applicant mixes, and how all that intersects with increasing RD applications per applicant (particularly for the applicants these colleges seriously consider), and the resulting decline in yield predictability.
But my impression is also that if you get outside of the most popular markets, this is far less true. Meaning RD admissions are still reasonably predictable overall, understanding holistic review creates an unobserved dimension that basic numbers cannot capture, but also understanding certain colleges are regularly looking for certain types of students such that a good fit improves your chances. In other words, in these not-hyper-popular markets, the traditional likely/target/reach approach, thoughtfully applied, will tend to work out well, even if you donāt know for sure what will happen at each specific college. That impression was based on me āstudyingā the available information going into S24ās application cycle, and his decision pattern so far is right on track with that impression as well, as are the other decisions I am hearing about.
So, yes, if you look only at colleges in or near Boston, in California, and so on, I think it is a very challenging situation from an RD prediction standpoint.
But if you are looking at, say, colleges in the Great Lakes/Midwest? Maybe with just a few exceptions (like Chicago, Northwestern, and maybe WUSTL (and even then I feel like I kinda get WUSTL)), I think it usually is more understandable. Which doesnāt mean there will be no āanomaliesā, but fewer, and those will often have an ascertainable basis (like, if you account for residency, school/major, or something like that).