So what does it say that I only got 9 out of 20 right on the elementary algebra section?
What is good about the ACCUPLACER (based on a cursory examination by an English teacher) is that Will's performance on the Elementary Algebra test should map straightforwardly to which Elementary Algebra topics he does and does not recall. A decent teacher, or even a good tutor or computer program should be able to bring him back up to speed quickly. College-Level Mathematics might take a lot longer, but what it is measuring is still fairly straightforward.
One does not get that feeling from the 11th grade NECAP math. Just as one example, if you miss a problem that requires you apply both probability and understanding of prime numbers, it isn't clear if you have a problem with probability, prime numbers, combining multiple math topics to solve one problem, etc. Especially if your problem is fluent application of multiple math skills in an authentic context, that's not something you can quickly brush up on.
Figuring out why a student got a low score in the NECAP would probably require a second diagnostic test, but I doubt such a test exists. As Tom Sgouros has been pointing out, the design of the NECAP is to rank students, not create a detailed profile of what they do an do not know.
A lot of this gets back to the fundamental question of what the purpose of a required math graduation test that's independent of from any specific course is. I would argue if that if you must do it, it should be a well-understood, transparent and specifically defined final check. The NECAP is just too idiosyncratic.