How to Write a Quadratic Function in Standard Form From a Graph: Three Equations From Three Points, Worked

Standard form is y = ax² + bx + c, and you can recover it from any graph that shows three identifiable points on the parabola. Read the coordinates, substitute each into y = ax² + bx + c, and solve the resulting three-equation system for a, b, and c. If the vertex is labeled instead, start from vertex form and expand. One condition governs everything: the three points must be distinct and must not lie on a straight line, or the system has no unique solution.

Identify the Three Facts a Parabola Graph Can Give You

Every graph hands you one of three problem shapes, and naming the shape before you touch a pencil decides which method runs. Three ordinary points on the curve means a three-by-three linear system. A labeled vertex plus one other point means vertex form with a single unknown. Visible x-intercepts plus one point means factored form. Physically inspect the curve across its whole visible width before committing, because a graph that shows only one arm of the parabola is still solvable but far more error-prone.

Vertices are the cheapest fact to use. Standard form hides the vertex, but vertex form y = a(x − h)² + k displays it as (h, k) directly, and you only have to solve for a. The catch is that textbook and exam graphs label the vertex roughly half the time; when they don't, you have to estimate it from symmetry, which introduces rounding error you'll carry into every coefficient.

Intercepts are the second cheapest when the parabola crosses cleanly at integer values. A graph crossing at (−2, 0) and (4, 0) gives you y = a(x + 2)(x − 4) with one unknown. Fractional or irrational intercepts — roots at values like 1.37 or 2 ± √3 — make this route worse than a plain three-point system, since you'll be multiplying messy decimals by hand.

The hard constraint is collinearity. Take three points from the same arm of a steep parabola close together and the system becomes ill-conditioned: your answer will technically solve it but drift badly from the true curve because the determinant of the coefficient matrix is near zero. Spread your chosen points across the full visible width, ideally one on each side of the axis of symmetry and one near the vertex.

Read Coordinates Off the Graph Without Guessing

Grid lines are your ruler, and the scale printed in the corner or implied by labeled tick marks — 1 unit per square, 2 units per square, 0.5 units per square — determines whether a point is exact. A point sitting dead on a grid intersection at (3, −4) is a fact. A point floating between lines at roughly (2.9, −3.8) is an estimate, and you should label it as one in your work.

Use the axis of symmetry as an error check. Every parabola satisfies x = −b/(2a), and visually the curve is a mirror about a vertical line through the vertex. If you read points at (0, 5) and (6, 5), the axis sits at x = 3, meaning the vertex's x-coordinate is 3 whether or not it's labeled. Hunting for symmetric pairs like this converts sloppy readings into exact ones surprisingly often.

Watch for truncated graphs. Many textbook figures show the parabola only between y = −10 and y = 10, and any intercept or vertex outside that window is simply absent. In that case, don't extrapolate — pick three points that are actually inside the frame and accept the longer system. Assuming an unshown y-intercept of 0 is the single most common error in this whole exercise.

Label each point with a letter before substituting. Writing A(−1, 2), B(2, −7), C(4, −3) at the top of your page costs ten seconds and prevents the classic disaster of swapping an x with a y halfway through the algebra.

Solve for a, b, and c From Any Three Points

Here's the full method with real numbers. Suppose the graph passes through A(−1, 6), B(1, 0), and C(3, 4).

1: Write the general model: y = ax² + bx + c. 2: Substitute A: 6 = a(1) + b(−1) + c, which simplifies to a − b + c = 6. 3: Substitute B: 0 = a(1) + b(1) + c, which simplifies to a + b + c = 0. 4: Substitute C: 4 = a(9) + b(3) + c, giving 9a + 3b + c = 4. 5: Subtract equation 2 from equation 1: −2b = 6, so b = −3. 6: Subtract equation 2 from equation 3: 8a + 2b = 4. Substitute b = −3: 8a − 6 = 4, so a = 1.25. 7: Back-substitute into equation 2: 1.25 − 3 + c = 0, so c = 1.75.

Check by re-substituting A: 1.25(1) + (−3)(−1) + 1.75 = 1.25 + 3 + 1.75 = 6. Correct. If a step fails the check, the arithmetic broke somewhere between step 2 and step 7 — usually a sign error on a negative x.

This subtraction trick works because two of the three equations share the same c. Pick the two equations closest in x-value and eliminate, then pair the remaining one with either. The full elimination method — multiply one equation so a coefficient matches, then subtract — always works, but the shortcut saves you from matrix machinery when the points are friendly.

Use the Vertex When the Graph Labels It

Vertex form is y = a(x − h)² + k, where (h, k) is the vertex, and converting to standard form means expanding: y = ax² − 2ahx + ah² + k, so b = −2ah and c = ah² + k. That expansion is the whole conversion, and it means you only need one additional point to pin a.

Take a parabola with vertex (2, −1) passing through (0, 7). Substitute: 7 = a(0 − 2)² + (−1), so 7 = 4a − 1, giving a = 2. Now expand: y = 2(x − 2)² − 1 = 2(x² − 4x + 4) − 1 = 2x² − 8x + 7. Verify the vertex from standard form: x = −b/(2a) = 8/4 = 2, and y = 2(4) − 16 + 7 = −1. Both match.

The failure case worth knowing: when the labeled vertex is itself an estimate, expand anyway, but then re-check against a third point. If the third point misses the curve by more than a rounding gap, the vertex label was wrong or you read it wrong, and the three-point system is the safer recovery route.

Vertex form also tells you the sign of a instantly. A parabola opening upward means a > 0 in every form, and a label like "(minimum)" beside the vertex confirms it. Downward, a < 0. Checking your final a against the visible opening direction is a one-second sanity test that catches sign flips.

Convert Between Vertex, Factored, and Standard Form

Three forms describe the same curve, and each graph feature maps to exactly one of them. A comparison makes the trade-offs clear:

From factored form, expansion is mechanical: y = a(x − r₁)(x − r₂) becomes ax² − a(r₁ + r₂)x + ar₁r₂, so b = −a(r₁ + r₂) and c = ar₁r₂. Check with roots 1 and 4 and a = 2: y = 2(x − 1)(x − 4) = 2x² − 10x + 8, and the vertex sits at x = 2.5, y = 2(1.5)(−1.5) = −4.5, matching the standard-form vertex formula.

Standard form gives away only one feature at a glance — the y-intercept is (0, c). That's why graph problems rarely hand you standard form to start: it hides everything the picture shows. The conversion step is always the same direction, from whichever form the graph makes obvious into standard form.

When a graph is unscaled and unlabeled entirely, none of the three routes gives exact numbers. Estimate the vertex and one point to one decimal, run the vertex route, and state your answer as approximate. Claiming exactness there is a lie the grader will catch.

Work Through a Downward-Opening Example End to End

The sign-flip case is where most students lose marks, so here is a full downward parabola. The graph opens downward and passes through (−2, 1), (0, 5), and (4, −3).

Substitute (−2, 1): 4a − 2b + c = 1. Substitute (0, 5): c = 5. Substitute (4, −3): 16a + 4b + 5 = −3, so 16a + 4b = −8 and 4a + b = −2. And from the first equation with c = 5: 4a − 2b = −4, or 2a − b = −2. Add the two simplified equations: 6a = −4, so a = −2/3. That negative value matches the downward opening, which is the check you want. Then 2(−2/3) − b = −2 gives b = 2/3. So y = −(2/3)x² + (2/3)x + 5.

Verify with (4, −3): −(2/3)(16) + (2/3)(4) + 5 = −32/3 + 8/3 + 5 = −24/3 + 5 = −8 + 5 = −3. Correct. The leading coefficient −2/3 is shallow, which matches a wide, gently curving downward shape — a second visual check that holds up.

Notice the y-intercept came free. Any graph showing where the curve crosses the vertical axis hands you c with zero algebra, which is why (0, c) should always be one of your three points when it's visible. Skipping it and using three other points wastes an hour of work.

Avoid the Mistakes That Quietly Break the Equation

The most expensive error is collinear points. If a graph is nearly flat across a narrow window and you pick three points within that window, the system's determinant approaches zero, coefficients balloon, and small reading errors amplify into wildly wrong curves. Test for it: if the three points look like they could sit on a straight line, they're too close together.

Second is misreading the scale. A grid labeled 2 units per square turns a "halfway" point at (1.5, 2) into (3, 4) on the page, and every substituted value is off by a factor of two. Check one coordinate against a labeled tick mark before doing anything else.

Third is assuming a graph that passes through the origin has c = 0 without checking — occasionally a curve clips the corner of the frame without actually going through (0, 0). Fourth is arithmetic on negatives: squaring a negative x is fine, but −b when b is negative flips sign, and that's where the b = −3 in our worked example would become b = 3 in a careless hand.

A quick final failure scenario: three points, but two of them are the same point read twice. The system looks like three equations but delivers only two independent facts, and elimination will produce 0 = 0. Recount your points before panicking.

Check Your Standard-Form Answer Against the Graphed Curve

Four checks catch nearly every wrong answer in under a minute. First, the sign of a against the opening direction: upward means a > 0, downward means a < 0. Second, plug the x-coordinate of the visual vertex into −b/(2a); the answer must match the vertex's x-position on the graph.

Third, re-substitute each of your original three points into the final equation and confirm all three come out exact. Fourth, check the y-intercept: c must equal where the curve crosses the y-axis, or (0, c) must be consistent with your readings. If any check fails, redo the substitution for the point that failed — the mistake lives there.

For decimal readings, allow a small tolerance. Points read to one decimal place legitimately produce coefficients that miss the true curve by a few hundredths, and a re-substitution off by 0.03 is not an error. Re-substitutions off by whole units are.

A worked verification from the vertex-form example: y = 2x² − 8x + 7 evaluated at x = 0 gives 7, matching the stated point (0, 7); the vertex x = 2 matches; a = 2 > 0 matches the upward opening. Three checks in three lines, and the answer is confirmed.

Handle Graphs Without Labels or Full Detail

When a graph carries no numbered ticks, state your assumption and proceed. Choose the vertex and one cleanly readable point, assign the implied unit grid, and answer in terms of that unit. This is exactly how exam figures with unlabeled axes get written up, and graders accept a clearly stated assumption over a confidently wrong exact answer.

Horizontal parabolas — x = ay² + by + c — occasionally appear and break the usual method because the roles of x and y swap. The graph opens left or right instead of up or down, and your three points substitute y-values into y² and y terms. Same system, same elimination, but a positive a now means opens right and negative means left.

Piecewise and truncated views: a graph showing only the right arm still permits a solution from three points on that arm, but the collinearity risk rises sharply there because the curve bends slowly over a small window. Pick the widest spread you can and include the visible vertex if one shows.

If the graph is a photograph of a handwritten sketch, expect noise. Read each point to the nearest half-unit, solve, and present the answer with a note that it's approximate. Perfect exactness from a sketch is not achievable and claiming it invites a challenge.

Turn the Standard-Form Equation Into the Graph's Features

The final conversion back is the payoff. From y = ax² + bx + c, the vertex sits at x = −b/(2a), y = c − b²/(4a), which is just the standard form of k = ah² + k after substituting h. The y-intercept is (0, c). The x-intercepts, if real, are x = (−b ± √(b² − 4ac))/(2a).

Take y = x² − 4x + 1. The vertex is at x = 2, y = 1 − 4 = −3. The y-intercept is (0, 1). The discriminant is 16 − 4 = 12, so the intercepts sit at x = 2 ± √3, roughly 0.27 and 3.73, which is why this parabola's graph usually shows awkward non-integer intercepts — exactly the case where the three-point system beats reading the intercepts directly.

This conversion also explains the shape. The leading coefficient controls width: |a| = 1 gives the parent parabola, |a| = 2 is twice as narrow, |a| = 0.5 is twice as wide. Comparing your recovered a to the graph's visible width is a free final check.

Every step here is reversible, and that round-trip — graph to equation to features — is the full exercise in standard form. Once you've done it three or four times with real coordinates, the method condenses into a single habit: read three facts, substitute, solve, verify.

FAQ

How many points do I actually need from the graph?

Three distinct, non-collinear points, because y = ax² + bx + c has three unknowns. A labeled vertex plus one point also works, since vertex form reduces the problem to one unknown. Two points are never enough unless the graph opens in a known direction and one is the vertex.

What if the three points I pick lie on a straight line?

Then the parabola is degenerate through those points and the system has no unique quadratic solution — the determinant of the coefficient matrix is zero. This happens when points cluster too tightly on a steep curve. Re-pick points spread across the full visible width, including one on each side of the vertex.

Can I get standard form with only the vertex and the y-intercept?

Yes, and it's the fastest route on a graph. Substitute (0, c) into y = a(x − h)² + k to solve for a directly, then expand. A parabola with vertex (3, −2) through (0, 7) gives a = 1, so the standard form is y = x² − 6x + 7 — three steps and no system.

Why does my answer open the wrong way compared to the picture?

Your leading coefficient has the wrong sign, which almost always means a sign error when substituting a negative x or negative b. Check the substitution for the point with the most negative coordinate. A downward-opening graph can never produce a positive a, so the flip is a guaranteed arithmetic bug, not an interpretation problem.

How exact should my coefficients be from a hand-drawn graph?

Read points to half a unit and round coefficients to two decimals, then state the answer as approximate. Exactness requires grid intersections; if the figure shows points on intersections, your answer should be exact fractions or integers, and a decimal that doesn't terminate signals a reading error.

Is standard form the same as vertex form?

No. Standard form is y = ax² + bx + c and reveals the y-intercept at (0, c); vertex form is y = a(x − h)² + k and reveals the vertex at (h, k). They describe identical parabolas, and converting between them is a matter of expanding or completing the square.

What if the graph shows only one x-intercept labeled?

One real intercept means the other root is also real and equal only if the curve is tangent to the axis at that point; otherwise it's hidden or fractional. Pick two additional points inside the frame and run the three-point system — a single intercept is one fact, not enough on its own.

How do I check the answer against the vertex formula?

Compute −b/(2a) from your final coefficients and compare it to the vertex's x-position on the graph. If they match, and the y-value c − b²/(4a) matches the vertex's height, the equation is almost certainly right before you even re-test the original points.