Look at a straight line. It goes up, or it goes down, but it never changes its mind — no bends, no bumps, no place where the graph swings from rising to falling. Now push the degree up by one. A parabola appears, with one turning point at the vertex. Push up again: the cubic bends twice, up to a local maximum and down to a local minimum. Quartic: three bends. Each additional degree buys the graph one more potential wiggle.
A polynomial of degree n can have at most n - 1 turning points. The widget below lets you slide the degree from 1 to 6 and watch the curve gain one more bend every time the degree clicks up.
The widget
Each position of the slider picks a canonical polynomial of that degree, chosen so the turning points are visible in the viewing window. The title bar above the curve tells you the formula and the count.
Try these
Walk the slider through every value and count the bends yourself.
- d = 1: y = x. A straight diagonal line. Zero turning points, because 1 - 1 = 0.
- d = 2: y = x^2 - 2. A parabola. One turning point, the vertex at x = 0.
- d = 3: y = x^3 - 3x. A cubic with two turning points — a local maximum at x = -1 and a local minimum at x = 1. The S-shape uses the full budget of n - 1 = 2 wiggles.
- d = 4: y = x^4 - 4x^2 + 2. A W-shape quartic. Three turning points — local min at x = -\sqrt{2}, local max at x = 0, local min at x = \sqrt{2}.
- d = 5: y = x^5 - 5x^3 + 3x. Four turning points, threading between five zeros along the x-axis.
- d = 6: y = x^6 - 6x^4 + 5x^2 - 1. Five turning points crammed into the same window.
Push the slider back to 1 and watch the sequence collapse: five bends, four, three, two, one, none. The number of wiggles is always exactly one less than the degree, for these particular polynomials.
The rule — degree n has at most n - 1 turning points
Why does the count stop at n - 1? The argument uses one fact from calculus, and it is worth knowing even if you have not met calculus yet.
A turning point is a place where the slope switches sign — the graph stops rising and starts falling, or vice versa. At the turning point itself, the slope of the tangent line is exactly zero.
The slope of a polynomial p(x) at each point is another polynomial called its derivative, written p'(x). The key fact: if p(x) has degree n, then p'(x) has degree n - 1. Differentiation knocks the degree down by one.
The turning points of p are the zeros of p'. From Polynomials — Introduction: a polynomial of degree k has at most k real zeros. Apply this to p': it has degree n - 1, so at most n - 1 real zeros. Hence p has at most n - 1 turning points. The bound is arithmetic on degrees, nothing more.
Why "at most" — sometimes fewer
The rule says at most, not exactly. Some polynomials of degree n have fewer than n - 1 turning points.
- y = x^3 has zero turning points. Its derivative is 3x^2, which equals zero only at x = 0. But there the slope does not switch sign; it touches zero and the curve keeps rising. This is an inflection point with horizontal tangent, not a turning point. Degree 3, zero wiggles.
- y = x^4 has one turning point. Derivative 4x^3, zero only at x = 0. The slope switches sign there, so it counts. One bend, even though the rule allows up to three.
The bound n - 1 is tight only when the derivative has all distinct real zeros. When the derivative has repeated zeros or complex zeros, the count drops. The polynomials in the widget were handpicked to hit the bound exactly.
End behaviour at each degree
As you slide the degree, notice the ends of the curve too. A pattern emerges.
- Odd degree, positive leading coefficient (degrees 1, 3, 5 in the widget). The curve runs from -\infty on the far left to +\infty on the far right.
- Odd degree, negative leading coefficient. The mirror image: +\infty on the left, -\infty on the right.
- Even degree, positive leading coefficient (degrees 2, 4, 6). Both ends rise to +\infty. Global U-shape, wiggles in the middle.
- Even degree, negative leading coefficient. Both ends sink to -\infty. Upside-down U.
The leading term dominates when |x| is large, because x^n grows faster than every lower power combined. End behaviour depends only on the degree and the sign of the leading coefficient.
The shape story
It is tempting to think "higher degree means more curved." The right intuition is higher degree means more places it can curve. A cubic does not bend harder than a parabola — it bends in one more place. Each additional degree is a licence for one additional wiggle; the polynomial is not obligated to use that licence, but it cannot exceed it.
This is why a degree-n polynomial can pass through n + 1 arbitrary points in the plane, a property called polynomial interpolation. It is the standard tool for curve fitting in numerical analysis, computer graphics, and scientific computing. ISRO's trajectory-planning software uses polynomial interpolation to smooth the path of a satellite between sparse data points. The reason polynomials work for this is the rule you just saw: increasing the degree increases the bending budget.
Why this matters
When you sketch a polynomial for an exam, the degree tells you how much complexity to expect. A degree-5 problem might have up to four wiggles, and if your sketch shows only two, you are probably missing something. If a sketch shows five wiggles for a degree-4 polynomial, something has gone wrong — the upper bound is three.
The same count shows up in optimisation: to find the highest or lowest value of a polynomial over an interval, you search among its turning points (at most n - 1 of them) plus the endpoints. It is one of the quiet constants of polynomial life.
Closing
Slide the slider. Count the wiggles. The rule "degree n has at most n - 1 turning points" is not a separate fact to memorise — it falls directly out of one observation (turning points are zeros of the derivative) and one degree-count (the derivative has degree one less). Simple picture, fundamental property.