I'll absolutely second this. I know kids who went to tons of prep classes for literally years. I had a friend whose parents paid for her to attend SAT prep summer camp for about a month every summer starting in sixth grade. Most of them did fine, but their scores were commensurate with that you'd expect from their report cards. I didn't study much and did very well. The students who did very well were mostly either those who were high-achieving in other respects or those who worked very, very hard. E.g. I know someone who wasn't particularly academically outstanding but got a perfect score with study, practice books only, no expensive classes.

How many times did you take it? If it was only twice, how can you be sure how much of that 100 point discrepancy was attributable to practice, and how much of it was essentially down to chance? I doubt you'd get the same score twice if you took it twice without studying in between.

I am curious if it can change now with so many good online educational resources.

Yeah, I also went to $GOOD_UNIVERSITY (in the UK). There were a lot of private high school students, but what stood out to me even more was the level of resourcing and support that the state schools that some students had gone to had given them.

A lot of these "good state-funded schools" pushed people to specific A-levels that looked good to universities, and had coaching on the university interviews and entrance exams (not to mention a much higher standard of teaching). These schools regularly had 100+ students a year get into $GOOD_UNIVERSITY. In contrast to my school that 2 people go in my year and none some years, and and couldn't even work out which entrance exams I needed to be entered for (which they had to do on my behalf).

I don't agree that the SAT is helpful though. These kind of standardised tests measure ability to prep far more than they measure intelligence. In my experience the universities are actually putting quite a lot of effort into making their courses widely accessible (setting lower entrance grades for people from less good schools for example). The problem is really with vastly uneven quality of teaching at the high school level which means that some students really are a lot more prepared for the courses than others (even if that doesn't reflect base level intelligence).

Or STEM schools (presumably not all) where ‘mastery’ is the tracked objective, achieved when you scored 80% or higher on tests. However, when mastery is translated to standard scoring models/transcripts it’s recorded as 100%.