Seeing as how I haven't checked in in awhile, hello everything. Special welcome back to Teach. So I've been working on some patterns, and have reduced my work into one, fairly underwhelming formula.
N = (D-A)^2+E / (2*A).
I'll be extra salty if this was stated somewhere else, other than the known "X^2+E = 2NA". I'm using the form D-A instead of X, because this is the fruit of working on the X+N square, and my observation that it seems like an overly complex way of iterating over A.
Here in this formula, if (2A) evenly divides (D-A)^2+E, A is valid, thus X, and you get N. Which isn't useful except, finding the intercept of these two equations, (where (D-A)^2+E < 2A) gives us the upper bound for A, (i.e, the lower bound for X). It may also be possible to push X into the negative (where B and A become swapped) and give us the upper bound for B.
There's some more patterns with this I'm exploring, and I could do a visualization, but what's happening here is fairly complex to show without coloring it manually.