M
Math Coach Amy
Amy Ferguson Moncure

Algebra 1 · Unit 10

Patterns That Are Not Linear or Quadratic

Describe scatterplots that need a piecewise rule or another curve, not a single line or parabola.

Jump to section:

What to Remember

Not a line, not a parabola

Some clouds rise then flatten, drop then rise, or follow two different trends on two sides of a cutoff.

Piecewise

One rule for x less than a number, another rule after. Example: a fee that is flat, then a per-mile rate.

Describe in pieces

Say what happens on the left, what happens on the right, and where it changes. Do not force one quadratic through a corner.

Other curves

Exponential growth, leveling off, or a U that is not symmetric can all appear. Name the shape you see.

Level 1

Easy

Name the shape: two trends, or flattening.

Example
Describe the pattern.

For small x, y stays near 2 (flat).

After about x = 3, y rises linearly.

A piecewise description fits better than one line or one parabola.

Answer: flat, then rising

1
A plot of hours of practice vs score rises quickly, then almost levels off. Why is a single line a poor model?
Show Answer

A line would keep rising. The data flatten. Describe two phases, or a leveling curve.

Answer: it flattens

2
Taxi fare is 4 dollars for any ride up to 1 mile, then 2 dollars per extra mile. Sketch the form (in words).
Show Answer

Horizontal from 0 to 1 mile at y = 4, then a ray with slope 2.

Piecewise linear.

Answer: flat then slope 2

3
Is a V-shape (down then up) a single quadratic opening up?
Show Answer

A parabola is smooth and U-shaped. A sharp V is two lines, piecewise.

Answer: piecewise V, not a parabola

4
Points follow y = 2ˣ for x = 0 to 4, not a parabola. What family?
Show Answer

Exponential.

Answer: exponential

5
A scatterplot is two separate tight lines, one for morning and one for afternoon. What should you report?
Show Answer

Two groups. A piecewise or two-plot description, not one cloud.

Answer: two groups

Level 2

Medium

Match a story to a piecewise or leveling shape.

Example
Phone data: speed is roughly constant, then drops after a cap. Which description matches?

Nearly horizontal, then a drop or a shallower slope after the cap.

Piecewise: one rate before the cap, another after.

Answer: piecewise at the cap

1
Describe.
Show Answer

Rises quickly, then levels off near 8.

Not linear; not a symmetric parabola.

Answer: rises then levels

2
A parking garage charges 3 dollars per hour for the first 2 hours, then 1 dollar per hour. Why is y = 3x a bad model for 5 hours?
Show Answer

After 2 hours the rate changes. Piecewise: 3x for 0 ≤ x ≤ 2, then 6 + 1(x − 2) after.

Answer: rate changes at 2 hours

3
A plant’s height vs days: slow, then fast, then slow again. Name a better summary than “quadratic.”
Show Answer

An S-shape (slow–fast–slow). A parabola would come back down; this may not.

Answer: S-shape / leveling

4
Scores vs hours of sleep: rise up to 8 hours, then a slight drop. Why might two pieces beat one downward parabola?
Show Answer

The peak is a change of story (enough sleep vs too much), not a symmetric U. Describe increase then decrease separately.

Answer: two pieces around 8 hours

5
A plot is linear for x < 10 and exponential-looking after a promotion at week 10. How do you describe it?
Show Answer

Piecewise: linear before week 10, a different growth after.

One parabola through both parts would miss the corner.

Answer: piecewise at week 10

Level 3

Hard

Choose among linear, quadratic, exponential, and piecewise.

Example
A scatterplot of weeks vs users is slow, then nearly a straight steep climb, with a corner. Compare a quadratic fit to a piecewise linear description.

A quadratic can look similar on a short window but will eventually bend the wrong way.

Piecewise linear names the corner (a launch week) and keeps a constant rate after.

If the story is “after launch we added a constant number each week,” piecewise linear matches the mechanism.

Answer: piecewise if there is a corner

1
Data: y ≈ 1, 2, 4, 8, 8, 8. Which combination of families?
Show Answer

Exponential doubling, then a cap (piecewise: exponential then constant).

Answer: exponential then flat

2
A U-shape that is much steeper on the right than the left. Why might “one quadratic” be incomplete?
Show Answer

A quadratic is symmetric about the vertex. Unequal sides suggest two different pieces or a different curve.

Answer: not symmetric

3
Gym fee: 20 dollars to join, then 5 dollars per visit, but visits 1–3 are free. Write a piecewise description of total cost vs visits n = 0, 1, 2, …
Show Answer

n = 0: 20.

n = 1, 2, 3: still 20 (free visits).

n ≥ 4: 20 + 5(n − 3).

Answer: 20, then 20 + 5(n − 3) after 3 visits

4
A residual plot from a linear fit shows a clear U. Next step?
Show Answer

Do not stop at the line. Try a quadratic, or split the domain if the U is a corner plus two trends.

Answer: try a curve or pieces

5
Why can two linear pieces be a better model than a quadratic for a tax bracket or a shipping rate that jumps at a cutoff?
Show Answer

The real rule is defined in pieces at a cutoff. A parabola is smooth and would invent a curve the company does not use.

Answer: the rule is already piecewise

Super Challenge

Super Challenge

1
Invent a real-world scatterplot that is neither linear nor quadratic, and describe it in two sentences.
Show Answer

Example: download speed vs time of day — high at night, low at 8 p.m., not a parabola.

A periodic or piecewise daily pattern fits better than one U.

Answer: e.g. time-of-day pattern

2
A student fits y = ax² + bx + c to a piecewise-linear fee and gets a decent R². Why can that still be the wrong model?
Show Answer

A parabola can pass near a corner on a short window.

The fee’s mechanism is two lines. Residuals may still show a corner, and predictions past the window will curve.

Answer: wrong mechanism

3
Give a checklist: when to say linear, quadratic, exponential, or piecewise after looking at a scatterplot.
Show Answer

Linear: straight cloud, constant first differences if binned.

Quadratic: U or hill, constant second differences.

Exponential: J-shape, roughly constant ratios.

Piecewise: a corner, a cap, or two groups with a cutoff in x.

Answer: see checklist

Need help with nonlinear scatterplots?

Book a Session with Amy