Once you know a parent function's shape, you almost never need to plot a transformed version from scratch. Every transformation — sliding it, flipping it, squishing it, stretching it — can be read directly off the equation , where is the parent function and , , , are numbers you control. Think of it as editing a photo: the picture (the parent graph) stays the same, and four sliders resize it, flip it and move it. The payoff is huge: one rule works for lines, parabolas, square roots, reciprocals, absolute values, and later for exponential and trigonometric functions too.
What each parameter does
| Parameter | Effect | Details |
|---|---|---|
| a | Vertical stretch/compression, and reflection | Multiplies all -values by ; stretch if , compress if ; reflects in the -axis if . |
| k | Horizontal stretch/compression, and reflection | Multiplies all -values by ; compress if , stretch if ; reflects in the -axis if . |
| d | Horizontal translation | Shifts the graph d units right if d > 0, or |d| units left if d < 0. |
| c | Vertical translation | Shifts the graph c units up if c > 0, or |c| units down if c < 0. |
The parameters outside the function ( and ) act on the output, so they behave the way you'd expect: triples every height, lifts everything 5 units. The parameters inside ( and ) act on the input before sees it, and that makes them feel backwards. Take . The parent's vertex happens when the input to the squaring is ; now that input is , which is when . So the vertex moves to — right 4, even though the equation shows a minus sign. Likewise in , the value that used to happen at now happens at , because : the graph is squeezed toward the -axis by a factor of . In general, to reproduce a parent input , the new graph needs , which solves to .
Mapping notation
The cleanest way to transform a set of known points is the mapping rule: every point on the parent graph moves to a new point on the transformed graph according to
The -coordinate is divided by then shifted by ; the -coordinate is multiplied by then shifted by . This works for any parent function — you never need the actual equation of the curve, just its key points. A reliable routine: list three to five key points of the parent in a table, add two columns for the images, apply the rule, plot, and join with the parent's shape. Special points stay special: the vertex of a parabola maps to the new vertex, the starting point of maps to the new starting point, and asymptotes shift with and .
Reading the equation carefully
The form only gives the right if has been factored out of the bracket. In , it is tempting to say "shift 6 right," but first factor: , so and — a horizontal compression by and a shift of only 3 right. Check with the starting point: at , confirming the graph starts at . The same care applies to signs: means , so (left 5), and means and . When an equation arrives in an unusual shape, rearrange it into the standard form before reading anything: is , and is , a reflection in the -axis followed by a shift up 5.
- Reflection in the -axis: is negative (flips up/down).
- Reflection in the -axis: is negative (flips left/right).
- A horizontal stretch by factor 2 needs , not (that would be a compression) — always think "."
- Points on the line of reflection do not move: under the -intercepts stay put; under the -intercept stays put.
Transforming the domain and range
Because the mapping rule moves every point, it moves the domain and range too. Horizontal parameters (, ) change the domain; vertical parameters (, ) change the range. For , the parent starts at , which maps to . The domain becomes . Since , the graph heads down from its start, so the range becomes . For the domain stays all real numbers (horizontal moves can't create gaps in a function defined everywhere) and the range becomes , read straight from the new vertex . Asymptotes behave the same way: for , the parent's asymptotes and shift to and , so the domain is and the range is . Stretches and reflections never move an asymptote that passes through the origin, which is why only and matter here. On a test, finding the new "anchor" (vertex, starting point or asymptote intersection) with the mapping rule, and then asking "which way does the graph go from here?", answers most domain and range questions in two lines.