Graph the boundary line for each inequality, use a solid line for ≤ or ≥ and a dashed line for < or >, shade the side that makes the inequality true, then look for the region where every shaded area overlaps. That overlap is the solution set, and a single test point decides each side. The method below runs in that order, because plotting all boundaries before shading anything prevents the most common error: erasing and re-drawing a half-plane because one line was drawn with the wrong style.
Work the Six Steps in This Order
Before picking up a pencil, rewrite every inequality in slope-intercept form, y = mx + b, so each boundary line's slope and y-intercept can be read directly. This step is diagnostic rather than cosmetic: an inequality like 3x − 2y > 6 becomes y < 1.5x − 3 only after dividing by −2, which flips the inequality sign. Skip the flip and every later shading decision inherits the error.
1: Rewrite each inequality with y isolated. Flip the inequality symbol whenever you multiply or divide both sides by a negative number. 2: Plot each boundary line. Graph it as if it were an equation, using the slope and y-intercept you just read off. 3: Decide solid or dashed. Solid for ≤ and ≥, dashed for < and >. 4: Pick a test point off the line, usually the origin (0, 0). Substitute it into the original inequality. 5: Shade the half-plane that makes the substitution true. If the origin sits on the line, use (1, 0) or (0, 1) instead. 6: Shade darker where all half-planes overlap, then label the corner points of that overlap region.
A worked example using y ≤ 2x + 2 and the region y ≥ x/2 − 1: the first boundary catches the test point (0, 0) because 0 ≤ 2 is true, so the region below that line is in. The second: 0 ≥ −1 is true, so the region above −1 is in. The overlap is the wedge between them, opening to the right.
When a test substitution produces a false statement, shade the opposite side rather than starting over; the boundary remains valid and only the half-plane flips. If two boundary lines are parallel and the shaded sides point away from each other, no overlap exists and the system has no solution — a legitimate, completable answer, not a dead end.
Shading Rules Take a New Student to a Correct Region
Shading is where most graphing errors live, not plotting. The standard convention treats the shaded region as "allowed" and the unshaded region as "excluded," which is why teachers say the answer is the double-shaded area. A statutory-style statement of the rule is worth writing on scrap paper: for y > mx + b, shade above; for y < mx + b, shade below; for y ≥ mx + b or y ≤ mx + b, shade the same side but draw a solid line.
| Symbol on the boundary | Line style | Shading test | Simple rule of thumb |
|---|---|---|---|
| y > mx + b | Dashed | Above | Strict, boundary excluded |
| y < mx + b | Dashed | Below | Strict, boundary excluded |
| y ≥ mx + b | Solid | Above | Inclusive, boundary included |
| y ≤ mx + b | Solid | Below | Inclusive, boundary included |
Two special cases trip up students who rely on the slope-intercept shortcut. The first is the horizontal line y = −3: the inequality y < −3 shades below the dashed line y = −3, and there is no slope to plot, so graph it as a horizontal segment at that height. The second is the vertical line x = 4: since x is not y, the test-point decision comes from substituting an x-value — for x ≥ 4, the origin gives 0 ≥ 4, which is false, so shade to the right of the solid line.
A vertical inequality in standard form like 2x − 3y < 9 needs the y-isolation step even when x looks like the interesting variable. Isolating y produces y > (2/3)x − 3, dashed boundary, above the line. If you leave the inequality in x form and test the origin, you get 0 < 9, which is true, but you've only tested x — the substitution must be into the original inequality with both coordinates, not just one.
Solid or Dashed: The Boundary-Line Distinction
A solid line says points on the boundary are part of the solution. A dashed line says they are not. That is the entire distinction between ≤ and <, and between ≥ and >, and students who treat line style as decoration inevitably mark a boundary point as a solution when it isn't.
Test this with the point (0, 3) against the system y ≥ x + 1 and y < −x + 5. In the first inequality, 3 ≥ 1 is true, and the boundary of y ≥ x + 1 is solid, so (0, 3) counts. In the second, 3 < 5 is true, but the boundary of y < −x + 5 is dashed — and (0, 3) doesn't lie on that line at all, so this doesn't decide solidity. The point (0, 5) does: substitute into the second inequality gives 5 < 5, false, so a point on the dashed boundary of y < −x + 5 is not in the solution despite feeling "on the line."
For a graph-checking habit, pick a corner point of the overlap and verify it in every inequality before submitting. If any single inequality fails at that point, the corner is not a vertex of the feasible region and the shading is wrong somewhere.
In real optimization problems from linear programming, an unbounded region often appears — one that wins on the objective function but has no corner because the inequalities don't close it off. The solid/dashed distinction tells the optimizer whether a boundary point is admissible; making the boundary solid when the model meant strict can change the answer and even the practicality of the solution. Instructors at the college level (for example, in standard algebra texts from publishers like Pearson and Cengage) treat this as a modeling point, not a drawing convention: the symbol set encodes feasibility, and it should be honored in the drawing.
Testing the Origin Costs Two Seconds and Saves Redrawing
The origin is the default test point because its coordinates are zeros, which makes arithmetic trivial: 0 > 0 + 3 reduces to 0 > 3, false, and the answer is immediate. Students who substitute other points usually do so because the origin is inconvenient — either it's on the boundary or it produces a trivial truth that doesn't distinguish the sides.
If the boundary passes through the origin, use (1, 0) or (0, 1) instead. For y ≤ 2x, substituting (0, 0) gives 0 ≤ 0, true, but because the line runs through the origin, the substitution gives no information: every point on the line satisfies the inequality whether the line is the boundary or not. Use (4, 0) and get 0 ≤ 8, still true, so the shaded side is unchanged — but with y ≥ 2x, (4, 0) gives 0 ≥ 8, false, so the other side is shaded. The test point only distinguishes halves when it's strictly off the boundary.
Here is an original worked computation from the values on this page. Suppose the system is y ≤ 2x + 2 and y ≥ x/2 − 1, and the graph is on a 10×10 grid. The two boundary lines intersect where 2x + 2 = x/2 − 1, so (3/2)x = −3, meaning x = −2 and y = −2. That gives the vertex (−2, −2). The overlap runs to the right of that point, so the drawing occupies roughly a quarter of the plot, and a full-region shading from (−2, −2) to the right edge at x = 10 stretches 12 units across, which is why the wedge needs to be redrawn larger if it drifts off the graph paper.
Test-point selection is also a hedge against sign errors elsewhere: if the answer feels wrong but every substitution checks, your test point was probably on the boundary. When an inequality is already in the form x ≥ c for a vertical boundary, the test point trick works with the x-coordinate alone; substituting a full point is still valid and produces the same side.
Where the Solution Lives: Reading the Shaded Region
The solution is the set of every point that satisfies all inequalities at once, and on a graph that means the darkest region — the intersection of all shaded half-planes. A point in a single shaded band but outside the overlap fails at least one inequality, so "it's in the shaded part" is only true when every band agrees.
For the system y ≤ 2x + 2 and y ≥ x/2 − 1, the region between the two lines and opening to the right is the answer. Test one interior point, say (0, 0): 0 ≤ 2 is true and 0 ≥ −1 is true, so it's in. Test an outer point like (0, 10): 10 ≤ 2 is false, so that point is out even though it looked between the lines on some sketches, because the upward branch of y = 2x + 2 sits above and to the left of the visible area.
Unbounded regions are normal in two-inequality systems, and asking whether a region is bounded becomes a meaningful question rather than a nitpick. In three or more inequalities, the overlap often forms a polygon and students are usually asked to name its vertices; for two inequalities the region is usually an open wedge or half-plane.
One subtlety when a "region" is drawn on a coordinate plane with axes: the axes themselves are not inequalities unless stated, and a point can lie in the fourth quadrant and still be a solution to the system. Plotting in the first quadrant only is a convention of many word problems, not a rule of graphing.
Two-Inequality Systems: Work Through a Complete Example
Take the system y < −x + 4 and y ≥ x − 2. The first boundary is a dashed line through (0, 4) with slope −1; the second is a solid line through (0, −2) with slope 1. Test the origin: 0 < 4 is true, so the region below −x + 4 is in; 0 ≥ −2 is true, so the region above x − 2 is in. The overlap is a wedge between the lines, its vertex where −x + 4 = x − 2, so 6 = 2x and x = 3, y = 1 — the vertex (3, 1). The vertex itself lies on the dashed boundary of the first line, so it is not part of the solution, even though it's the wedge's corner.
Now add a third inequality, y ≤ 3, to the same system. The boundary is the solid horizontal line y = 3, and substituting (0, 0) gives 0 ≤ 3, true, so the band below the horizontal line becomes part of the shading. The overlap of all three regions is now a triangle whose vertices are (3, 1), the intersection of y = 3 with y = −x + 4 at (1, 3), and the intersection of y = 3 with y = x − 2 at (5, 3). The region is bounded, which is a new qualitative fact—bounded systems come up in optimization as closed feasible regions where a maximum is guaranteed, whereas the two-inequality version's unbounded wedge needs a separate check.
Note the vertex (3, 1) has a dashed boundary and cannot be an optimum for the closed region, but the other two vertices, both on solid boundaries, can be. When the objective function is linear, an optimum for a bounded region occurs at a vertex, so the solid/dashed distinction changes which candidate points are admissible, not just which ones are drawn.
If a student graph is so small that the intersection of y = 3 and y = x − 2 at x = 5 sits at the paper's edge, the triangle appears truncated and the third vertex goes missing. Practice on a graph with a margin around the plotted region, sized to cover every named vertex.
Coverage of Every Sub-Region, Not Just the Corner Points
The shaded region includes points no line passes through, and checking only vertices misses that. For the region y ≥ x + 1 and y < −x + 5, an interior point like (0, 2) satisfies both (2 ≥ 1, 2 < 5), while the corner where the lines meet, at (2, 3), sits on the first line and is admissible there but not on the second line if the region's boundaries are dashed. The interior always satisfies every inequality, which is why one test point off the boundary is enough to verify shading but not enough to verify corner admissibility.
Students preparing a graph for an exam should shade rather than only outline the region, since a blank interior in a "shade the region" question loses points even when boundaries are correct. If the exam asks for the region defined by a system of four inequalities, filling in the overlap in a darker tone is the safe standard.
There is a graphing-calculator check for absolute confidence: plot each inequality with the TI-84's "inequality" shading feature or with an online system, and compare the quadrant pattern. If the calculator shades the opposite half-plane from your hand drawing, the error is almost always an arithmetic slip, not a conceptual one, and the test-point substitution catches it directly. Check both directions once with a positive-slope line and a negative-slope line before the first exam, rather than during it.
The frequent failure in textbook word problems is forgetting to clip the region to the first quadrant when a context (for example, hours on a machine) allows only non-negative values. A word problem about production capacity really includes x ≥ 0 and y ≥ 0, and the region then becomes a polygon clipped by the axes.
Region-Boundary Points: The Vertex Distinction in Practice
A corner where two lines cross is not automatically a vertex of the feasible region. Take a system with an inequality y ≥ 0.5x, drawn with a solid boundary, and y ≤ 1. The first corner is where y = 0.5x meets y = 1, at (2, 1). The point satisfies both inequalities—1 ≥ 1 is true and 1 ≤ 1 is true—so it is a legitimate vertex of the overlap. Now replace the first inequality with y > 0.5x and the same intersection point is still in the closed rectangle's corner but no longer admissible, because 1 > 1 is false.
The distinction matters when writing a system whose solution is a specified polygon. If the intended answer includes vertex (0, 0), both inequalities touching that point must be non-strict at that vertex, meaning the boundaries must use ≤ or ≥. If the intended answer excludes the boundary near the origin, at least one touching inequality must be strict, and the polygon is open at that corner. This is the mechanical rule behind what instructors describe as a "closed feasible region," and the calculator or online grapher will render the distinction with a solid or dotted curve.
A point exactly on a corner sometimes counts for one inequality and not the other: an angle of 90° between an included and excluded boundary is the standard visual for the feasible-region diagrams in linear programming texts, where objective lines are then graphed tangent to the corner. Labeling only corners misses the direction of the region, so the label should note whether the polygon is closed at each vertex.
Graphing-Manual and Calculator Verdict: When the Drawing Is Enough
The hand-graphed region is the examined answer in algebra, and the calculator is a spot-check for the shading decision. Hand graphs earn credit for the boundary style, the shading, and the label; the calculator earns none of those. Reach for the technology when a corner coordinate is awkward or when you're checking a 4-inequality problem with three intersections.
| Method | Best for | Fails at | Credit earned in an exam |
|---|---|---|---|
| Pencil, slope-intercept | 2–3 inequalities, integer vertices | Large grids, fractional corners | Full marks when boundaries and shading shown |
| Graphing calculator | Verifying a shading side | Showing dashed vs. solid clearly, labeling | Usually not accepted as sole answer |
| Algebra-only (no graph) | Proving no solution, finding exact vertices | Communicating the region visually | Depends on the question's instruction |
Pick pencil-and-slope for any exam question; use the device for a quick sign check when the region runs off the paper. One real drawback of the calculator check: its shading view often hides whether the boundary is solid or dashed at a vertex, so the very point that decides admissibility looks identical in the display—confirm the symbol on paper before trusting it. Skip the technology for a single inequality, where one test point does the job faster than anyone can enter three coefficients.
FAQ
What does "the region where the shaded areas overlap" actually mean when three inequalities shade in different directions?
It means every point inside the darkest area satisfies all three inequalities at once. For three shaded bands, a point can sit inside one or two bands and still fail the third, so the answer area is only the triple overlap. Shade each band lightly and darken the overlap so the region is unambiguous. If you notice a point in the overlap that fails a test, recheck the inequality that it fails—the error is virtually always a mis-signed slope.
Why can't I just plot the equations and shade alternate pieces?
Proving that graphing the three solution bands of alternating sides covers exactly three alternations is possible with a drawing, but the intersection behavior of half-planes in two variables isn't the same as parity. Draw the boundary lines first—y = 2x and y = −x + 3—and shade the sides each inequality allows as a half-plane, then darken the overlap. A point sitting on a single band but not the other is not a solution.
Is (0, 0) always the right test point?
No, and the exception matters. If any boundary line passes through the origin, substituting (0, 0) gives a true statement or a false one that's true for a reason that has nothing to do with the chosen side. Use (1, 0) or (0, 1) instead, which is off the line whenever the line isn't also horizontal or vertical through those points. Check the substitution produces a "clean" truth result rather than a boundary equality.
What's the difference between a "solution set" and a "solution"? The terms feel interchangeable.
A solution is a single ordered pair, but a solution set is every pair that satisfies all the inequalities; graphing displays the solution set as a region. For y ≥ 2 and y ≤ x, the point (3, 3) is a solution, while the set is the infinite triangle where y is between 2 and x. You can't verify a solution set by testing one point; you verify a point's membership in the set, then rely on the region's uniform interior for the rest.
Can a system of linear inequalities have no solution? The graph looked empty in class.
Yes, when the shaded half-planes don't intersect at all, which happens commonly with parallel boundaries in strict inequalities. y < x − 1 and y > x + 2 share the same slope but the first line is always below the second, so the two shaded sides point away from each other. The correct answer is "no solution," and the next step is to state that symbolically instead of forcing a region.
Do I have to graph in the first quadrant?
Only if the problem's context requires non-negative values, but not because of the inequality system itself. Hours in a word problem or quantities of a product can't be negative, so a context that says "hours" or "items" usually justifies adding x ≥ 0 and y ≥ 0. On a pure math problem, graph the full plane unless the quadrant restriction is stated.
Does solid or dashed come from the inequality or the equation's sign?
The inequality symbol, read left to right. A solid line comes from ≤ or ≥, and a dashed line from < or >, regardless of whether the left side is x or y. Flipping the inequality after multiplying by a negative flips both the symbol and thus the boundary style—the sign you end with is the sign that governs.