How To Find Equation Of Parabola

Three Points Are Enough

You don't need a graphing calculator or any fancy software to find the equation of a parabola. Give me three points and I can write the equation in maybe five minutes if I'm being careful, or ten if I'm tired and double-checking arithmetic. The standard form is y = ax² + bx + c, and plugging in each point gives you a system of three linear equations in three unknowns. Let's say your points are (1, 3), (2, 7), and (-1, -1). Substitute each one: a + b + c = 3, 4a + 2b + c = 7, and a - b + c = -1. Solve that system by elimination or substitution or whatever you were taught, and you get a = 1, b = 1, c = 1. The equation is y = x² + x + 1. Done.

How To Find Equation Of Parabola When You Know the Vertex

Sometimes you're given the vertex instead of three points. That's actually easier in most cases because the vertex form y = a(x - h)² + k collapses the problem down to finding a single unknown. Plug in the vertex (h, k) and one other point, then solve for a. That's it. No three-equation system to manage. I once had a student who mixed up the sign inside the parentheses when substituting the vertex. He wrote y = a(x + 3)² + 5 for a vertex at (-3, 5) instead of y = a(x + 3)² + 5 being correct and got the wrong answer on every subsequent step. The takeaway: the formula uses minus h, so if h is negative you end up adding. Write it out carefully before you plug numbers in. There's a trade-off between the two forms worth knowing. The standard form gives you the y-intercept directly — it's just c. The vertex form gives you the turning point immediately, which matters if you're sketching a graph or doing optimization. If you need both, you can convert between them by expanding or completing the square. Completing the square on y = 2x² - 8x + 5 gets you y = 2(x - 2)² - 3, so the vertex is at (2, -3). It's a mechanical process, not a hard one, but you have to be comfortable manipulating algebra quickly or you'll waste time on something simple.

Edge Cases That Trip People Up

Here's something most textbooks skip: not every quadratic-related curve that looks like a parabola opens up or down. If two of your three points share the same x-coordinate, you don't have a vertical parabola. You have a horizontal one, and the equation takes the form x = ay² + by + c. This comes up more often than you'd think in physics problems involving projectile motion with rotated frames, or in geometry where the axis of symmetry isn't parallel to the y-axis. My workaround is straightforward — I check for duplicate x-values first before setting up any system. If I see them, I swap the roles of x and y, solve the resulting system for the horizontal parabola, and I'm done. It saves me from chasing a nonexistent solution for an hour. Another pitfall: three collinear points. They won't give you a parabola, they'll give you a line, and the coefficient a will come out to zero. That's mathematically valid — a line is a degenerate parabola — but it's usually not what the problem is asking for. If your a value is zero or suspiciously close to zero, go back and verify your points aren't collinear.

A Note on Numerical Stability

When the points are very close together or nearly collinear, solving the system numerically can become unstable. The matrix involved in Cramer's rule or Gaussian elimination develops a small determinant, and rounding errors blow up. In practice this means you might get a = 1.0003 when the true answer is 1, or worse, a sign error that flips your entire parabola. If you're working with real data — experimental measurements, surveying coordinates, anything with noise — consider using a least-squares fit instead of exact interpolation. Three points force the parabola through every measurement, including the error. Four or more points with a regression approach gives you something that actually represents the underlying trend. Most spreadsheets and Python libraries handle this in a single function call, and it's genuinely faster than wrestling with a poorly conditioned system by hand. The exact formula still applies. The process of finding it hasn't changed in a few hundred years. What changes is knowing when the formula is the right tool and when it's better to step back and use something else.