Concavity Calculator

What Is Concavity?

Concavity describes which way a curve bends, independent of whether it rises or falls. A concave up stretch cups upward like a bowl (∪): the slope is increasing, and every tangent line sits below the curve. A concave down stretch caps downward (∩): slope decreasing, tangents above. The second derivative is the detector:

f″(x) > 0: concave up. f″(x) < 0: concave down. Sign change: an inflection point.

Concavity is the second layer of a curve's story. The first derivative says where you're going (up or down); the second says how the trip is changing (speeding up or easing off). A rising, concave-down graph is growth losing steam; rising and concave up is growth compounding. Markets, epidemics, and learning curves are all read through exactly this lens.

How to Use the Concavity Calculator

Enter a function. The calculator computes f″ symbolically, solves f″ = 0 numerically on the window [−6, 6], verifies a genuine sign change at each candidate (filtering out impostors like x⁴'s flat point), and reports the complete partition: every concave-up interval, every concave-down interval, and every inflection point with its coordinates. The graph marks the inflections in teal, at exactly the spots where the curve switches its bend.

Worked Example

f(x) = x³ − 3x². Second derivative: f″ = 6x − 6, zero at x = 1, and the sign genuinely flips there (negative before, positive after):

Concave down on (−∞, 1), concave up on (1, ∞), inflection point at (1, −2)

Reading the graph confirms it: the left arm caps over the local maximum at x = 0, then the curve straightens momentarily at (1, −2) and cups upward through the local minimum at x = 2. The inflection sits exactly midway between the two extremes, as it does for every cubic. The bell-curve chip is the applied classic: e^(−x²) inflects at x = ±1/√2, the points where the normal distribution's bend reverses, and (not coincidentally) exactly one standard deviation from the mean in its statistical costume.

The f″ = 0 Trap

The most-tested subtlety in this topic: f″ = 0 does not guarantee an inflection. The fourth chip, f(x) = x⁴, has f″ = 12x², which vanishes at 0 but never goes negative: the curve is concave up on both sides, and (0, 0) is a minimum, not an inflection. The definition requires a sign change, and this calculator checks it explicitly rather than trusting the equation. The converse trap also exists: an inflection can occur where f″ fails to exist (x^(1/3) at 0 flips concavity across a vertical tangent). The reliable procedure is the same as for critical points one derivative down: find candidates (f″ zero or undefined), then test the sign on each side.

Common Mistakes to Avoid

• Confusing concave up with increasing. They are independent: e^(−x) is decreasing yet concave up the whole way. Direction is f′'s job; bending is f″'s. Any of the four combinations can occur.
• Declaring inflections from f″ = 0 alone. Verify the sign change. x⁴ is the canonical counterexample, and graders plant it deliberately.
• Reporting inflection x-values without the y-coordinate. An inflection is a point on the curve: (1, −2), not "x = 1." Applications need the location, not just the input.
• Using concavity language for the wrong object. "The function is concave up" must specify where: most functions bend both ways on different intervals. Interval-by-interval reporting, as this tool practices, is the correct grammar.

Real-World Applications

Inflection points are where trends change character, and entire industries watch for them. Epidemiologists call the inflection of a case curve "the peak of daily growth": before it, each day outpaces the last; after it, growth decelerates even as totals climb. That single point drove policy worldwide in 2020. Business reads product adoption the same way: the S-curve's inflection marks the shift from accelerating to saturating sales, the moment marketing spend stops compounding. Economists date recessions ending not when GDP rises but when its decline slows, a concavity observation.

In engineering and statistics the bell curve's inflections at ±σ give the standard deviation a geometric meaning your eye can find on any histogram. Structural engineers track the inflection points of a loaded beam's deflection curve, where bending moment vanishes and reinforcement can be spliced. And in machine learning, loss curves are monitored for concavity changes that signal a training phase transition. Wherever a quantity's acceleration matters more than its level, the second derivative, and this page's analysis, is the instrument.

Frequently Asked Questions