In short
A graph is sitting on the page. Somewhere on it is a straight line. The question says "find the equation of this line." Your hand should already be doing three things in this exact order:
- Pick two clean points — two grid intersections (integer coordinates) that the line clearly passes through.
- Compute the slope: \quad m = \dfrac{y_2 - y_1}{x_2 - x_1}
- Plug into point-slope: \quad y - y_1 = m(x - x_1)
That is the whole recipe. Three steps, ninety seconds, equation on the page. The danger is not the algebra — it's reading the wrong points off the graph.
This is a recipe chapter for the graph-reading version of the line-equation question. A sibling article handles the case where the two points are handed to you in the question. Here, you are extracting the two points yourself by looking at a printed graph — and the failure mode is completely different. The slope formula and point-slope substitution are identical to the sibling recipe; what is new is the eyeball step at the start.
In CBSE Class 9 and Class 10 board papers, "the graph of a linear equation in two variables is shown — find the equation" is a standard 2-mark or 3-mark question. The marks aren't lost in the algebra. They're lost when a student squints at the graph, guesses that the line passes through (1.3, 2.7), and writes a wildly wrong equation off a misread point. This article is about not doing that.
What's a "clean" point
A clean point is a place where the line passes exactly through a grid intersection — both the x-coordinate and the y-coordinate are clearly integers (or at worst, clearly readable halves like 0.5). You can put your fingertip on it and be sure of both numbers.
A dirty point is anywhere the line is between gridlines. Maybe x = 1 and y is somewhere between 2 and 3 — but is it 2.4? 2.5? 2.6? You will guess, you will guess wrong, and your slope will be off.
Why this matters so much: the slope formula amplifies tiny reading errors. If you guess y = 2.5 when the true value is 2.4, that 0.1 unit slip can shift your slope from 2 to 1.85 — and now your whole equation is wrong. Picking grid intersections sidesteps the entire issue: integers are unambiguous.
Worked example 1 — line through the y-axis
Graph shows a line through $(0, 1)$ and $(2, 5)$
Step 1 — pick two clean points. Scan the line. It crosses the y-axis at y = 1 exactly — that's the point (0, 1). Two units to the right, it passes through the gridline crossing at height 5 — that's (2, 5). Both are unambiguous integer coordinates. Circle them mentally.
Step 2 — slope. Label (x_1, y_1) = (0, 1) and (x_2, y_2) = (2, 5).
Step 3 — point-slope. Use the friendly point (0, 1) — the zero in x_1 kills a bracket:
Sanity check with the spare point (2, 5): 2(2) + 1 = 5. Tick.
Why pick the y-axis crossing whenever it's clean: it makes x_1 = 0, which collapses point-slope into the slope-intercept form y = mx + c in one step. The graph is literally handing you the y-intercept, and the recipe rewards you for grabbing it.
Worked example 2 — line through the x-axis
Graph shows a line through $(-1, 0)$ and $(1, 4)$
Step 1 — pick two clean points. The line crosses the x-axis exactly at x = -1 — that's (-1, 0). Two units to the right, it passes through the gridline at height 4 — that's (1, 4). Both clean.
Step 2 — slope.
Step 3 — point-slope. Use (-1, 0) — the zero in y_1 kills the left bracket:
Sanity check with (1, 4): 2(1) + 2 = 4. Tick.
Why write (x - (-1)) before simplifying to (x + 1): skipping that intermediate is exactly where students drop a sign and write (x - 1), which gives y = 2x - 2 — a parallel line shifted four units down. Two extra seconds of writing the brackets explicitly is your insurance.
Worked example 3 — the trap (no clean points nearby)
Graph shows a line that crosses gridlines mid-square
The line is clearly drawn, but every time it crosses a vertical gridline it lands between horizontal gridlines. At x = 1 it looks like the line is somewhere around y \approx 1.7. At x = 2 it's around y \approx 2.4. You squint, you guess.
Don't. Don't pick those points. Keep scanning the line in both directions until you find two grid intersections where the line passes exactly through a corner of a grid square. They might be far apart — that is fine, far-apart clean points are better than nearby dirty ones. Suppose, after scanning, you find the line passes cleanly through (-3, 0) and (7, 7) — a wide span, but both unambiguous.
Step 2 — slope.
Step 3 — point-slope. Use (-3, 0):
The fractional slope 7/10 is exactly why those nearby points looked dirty — between any two adjacent gridlines, the line rises by 0.7 units, which never lands on a horizontal gridline within a single x-step. The far-apart pair was the only clean reading available, and it gave you the exact equation.
Why far-apart clean points are better than nearby dirty ones: errors in reading shrink in the slope formula when the denominator x_2 - x_1 is large. A 0.1-unit misread divided by a span of 10 shifts the slope by only 0.01. Divided by a span of 1, it shifts the slope by 0.1 — ten times worse. Spread out, then read carefully.
What if no clean points exist at all?
Sometimes — rarely, but it happens — the line on the printed graph genuinely doesn't pass through any grid intersection cleanly. Two escape hatches:
- Use the axis intercepts if they're given. A graph showing an x-intercept marked clearly at (4, 0) and a y-intercept marked clearly at (0, 3) gives you two clean points by definition. Use intercept form \dfrac{x}{a} + \dfrac{y}{b} = 1 to write \dfrac{x}{4} + \dfrac{y}{3} = 1 in one step.
- Read the slope and one intercept directly. If the y-intercept is clean and the slope is visible (one unit right, two units up = slope 2), write y = 2x + c and substitute the intercept.
In every case, the principle is the same: only ever read coordinates you are certain of. If you find yourself thinking "it's about 2.6" — stop, scan further along the line, find a true grid intersection.
The full recipe in one breath
- Scan the line until you find two grid intersections it passes exactly through.
- Compute m = (y_2 - y_1)/(x_2 - x_1).
- Drop m and one of the points into y - y_1 = m(x - x_1).
- Tidy to y = mx + c if the question asks for slope-intercept form.
- Verify with the spare point — plug it in, check it satisfies the equation.
Step 5 is the one students skip and lose marks on. It costs ten seconds. It catches every sign error and every misread coordinate.
Going deeper
For the curious
The "pick clean grid points" rule is a baby version of a much bigger idea in numerical mathematics — conditioning. A computation is well-conditioned when small input errors produce small output errors, and ill-conditioned when small input errors blow up.
The slope formula m = \Delta y / \Delta x is well-conditioned when \Delta x is large (far-apart points) and ill-conditioned when \Delta x is small (close-together points). In the language of error analysis, the relative error in m scales like 1/\Delta x for a fixed absolute error in y. That's the formal reason your physics teacher tells you to "use the longest possible base" when measuring slopes from experimental graphs in Class 11 lab work — the same principle you're applying here for free.
The same idea reappears in linear regression, where best-fit slope estimates have variance \sigma^2 / \sum(x_i - \bar{x})^2 — wider x-spread, lower variance, sharper slope. Picking far-apart clean points on a graph is the unweighted, two-point version of that whole machinery. Once you see it, the habit will follow you from CBSE Class 9 graphs all the way to JEE Advanced, undergraduate physics labs, and statistical inference.
References
- NCERT Class 9 Mathematics, Chapter 4: Linear Equations in Two Variables — graph-reading exercises.
- NCERT Class 10 Mathematics, Chapter 3: Pair of Linear Equations in Two Variables — graphical method.
- Khan Academy — Slope from a graph.
- Wikipedia — Linear equation: graph reading.
- Wikipedia — Condition number — why far-apart points give sharper slopes.