I tested out of multivariable calculus in college and to this day, it is one of my biggest regrets. My peers who took the course have an intuitive understanding of the material that I never developed. Sometimes there's a difference between learning on your own and working through the material with others.
For 99% of multivariable calc students, once the course is over and the social environment moved on, there is no reason or motivation to leaen it. Like most of college coursework, it's just a ritual.
There are loads of information available on this. For example, Khan Academy has a very nice multivar glass, done by Grant Sanderson (3Blue1Brown). Why are you limited by what happened (presumably) years ago?
The lack of homework, I think. It's not a natural thing to do. Apart from an explicitly pedagogical environment I have found the best thing to suggest people is to try to work their learning into a project. For math this can be hard unless you also have relevant engineering experience to make something with your math you are learning.
The first part of 18.02x (Multivariable Calculus) is open at the moment, not for certificate, but the autograder works. The second part starts next month
I'm still struggling to find the right blend of self study and group study. With peers you can coast along through the help of others to pinpoint you where you missed something, what's the "right" path to a solution. Which means .. you didn't really understand it. But sometimes I found solutions on my own, but then I realized that there were still some issues in them. I'd say it's important to self study but to reality check regularly.
