The optimum has an address, and it costs one line to find: $h = -\dfrac{b}{2a}$ — read aloud h equals minus b over two a. $h$ is the x-coordinate of the vertex, the input at which the model turns; $a$ is the coefficient of the squared term and $b$ the coefficient of the first-power term, both taken with their own signs. $c$ is not invited: sliding the entire model up or down cannot move the input that optimizes it.

Where does it come from? Next chapter's answer, one year early: the derivative of $ax^2 + bx + c$ is $2ax + b$, and a smooth curve turns exactly where its slope is zero. Set $2ax + b = 0$ and you get $x = -\dfrac{b}{2a}$ — the same address, derived rather than remembered. That is worth knowing now, because it means this formula is not a trick from algebra class; it is calculus already solved for you.

Two disciplines carry the marks. The minus belongs to the FORMULA, and b brings its own sign to the party as well — with $b = -96$, $-b$ is a cheerful $+96$. And the divisor is $2a$, never $a$: divide by a alone and you land on the sum of the roots, exactly twice as far out as the point midway between them.