Average rate of change: the interval matters
About the Author
Marko Šinko
Co-Founder & Lead Developer, AI Math Calculator
The average rate of change calculator evaluates [f(b) − f(a)] / (b − a). It measures net output change per input unit between two distinct inputs. For f(x) = x² from x = 1 to x = 3, the output rises from 1 to 9: the rate is (9 − 1)/(3 − 1) = 4. This does not mean the function rises by 4 at every point.
From endpoint values to a rate
First evaluate the function at both inputs. Subtract the outputs in the same order as the inputs, then divide. The secant line joins (a, f(a)) to (b, f(b)); its slope is this quotient. OpenStax’s treatment of rates of change develops the formula for functions, graphs, and tables and explains output units per input unit.
In the x² example, the secant passes through (1, 1) and (3, 9). Its equation is y − 1 = 4(x − 1), or y = 4x − 3. At x = 2 this line has height 5, although the curve has height 4. The straight connection summarizes endpoint change; it does not reconstruct the curve. For slope angles or line equations from general coordinates, use the slope calculator.
Swapping both endpoints leaves the result unchanged: (1 − 9)/(1 − 3) = (−8)/(−2) = 4. Reversing only the numerator gives the wrong sign. Equal inputs make the denominator zero, even if the outputs are also equal; the rate is undefined rather than zero.
Three intervals, three different messages
| Function and interval | Endpoint outputs | Average rate |
|---|---|---|
| x² on [1, 3] | 1 and 9 | 8/2 = 4 |
| x² on [−2, 2] | 4 and 4 | 0/4 = 0 |
| √x on [0, 4] | 0 and 2 | 2/4 = 1/2 |
The first example has a positive net change. In the second, x² falls from 4 to 0 and then climbs back to 4. The horizontal secant has slope zero, but the function is not constant. The third example is valid even though √x has no finite derivative at x = 0: the two endpoint outputs and their quotient still exist.
As another check, enter x^2-1/x with a = 2 and b = 4. The outputs are 7/2 and 63/4, so the change is 49/4 and the average rate is 49/8 = 6.125. Keeping fractions until the final division avoids unnecessary rounding.
Use units to interpret measured values
Suppose a tank contains 20 liters at minute 2 and 8 liters at minute 5. In known endpoint values mode, enter a = 2, b = 5, f(a) = 20, and f(b) = 8. The result is (8 − 20)/(5 − 2) = −4 liters per minute. The negative sign indicates a net loss as time increases; the two observations alone do not establish a constant draining rate.
For a table, use the two rows that bound the requested interval. Intermediate rows help describe what happened, but do not enter the endpoint quotient. If interval lengths differ, a plain average of adjacent rates can mislead. For readings (0, 0), (1, 10), and (4, 16), the adjacent rates are 10 and 2. Their unweighted mean is 6, while the whole-interval rate is 16/4 = 4. Weight by duration instead: (10 × 1 + 2 × 3)/4 = 4.
Change in position divided by elapsed time gives average velocity. An object that goes 3 meters away and 3 meters back in 6 seconds has zero average velocity and average speed 1 meter per second. Speed uses total distance traveled; endpoint positions alone cannot supply that distance. Likewise, this calculator’s rate is an absolute change per input unit, not a percentage growth rate.
Average rate versus instantaneous rate
Writing b = a + h turns the formula into [f(a + h) − f(a)]/h, the difference quotient for a nonzero step. If its limit exists as h tends to zero, that limit is f′(a), the instantaneous rate. OpenStax’s definition of the derivative explains this transition from a secant slope to a tangent slope.
For x² starting at a = 1, expand (1 + h)²: the quotient becomes (2h + h²)/h = 2 + h. At h = 2 the average is 4; at h = 0.5 it is 2.5; at h = 0.1 it is 2.1. The limit is 2. This page evaluates finite intervals. Use the derivative calculator for differentiation, or the tangent line calculator for a line at one point.
If a function is continuous on [a, b] and differentiable on (a, b), the mean value theorem guarantees an interior point with derivative equal to the average rate. For x² on [1, 3], 2c = 4 gives c = 2. The mean value theorem calculator addresses that additional task. An endpoint quotient by itself does not check the theorem’s hypotheses.
Average height is a different calculation
The average value of a function is its integral over the interval divided by the interval length. It measures average height, with the same units as the function’s output. On [1, 3], x² has average height 13/3 ≈ 4.333333333, while its average rate of change is 4. Use the average value of a function calculator when the question asks for mean height or an integral average.
Domain and precision checks before interpreting the answer
A numerical endpoint quotient requires finite, real values at two distinct inputs. Thus ln(x) with a = 0 fails, but 1/x between −1 and 1 has endpoint quotient 1 even though the function is undefined at zero. In the latter case, a secant can join the two points; it cannot establish an unbroken process through the interval. The tool warns when sampled interior values fail, and sampling cannot find every gap.
Use parentheses for negative bases: −x² means −(x²), whereas (−x)² is positive away from zero. Powers associate to the right, so 2^3^2 means 2^(3^2). Write multiplication explicitly, such as 2*x or 2*pi. The accepted function names, bounds, and real-number limits are listed with the controls; piecewise formulas, complex values, and symbolic endpoint variables are not supported.
Displayed decimals are rounded numerical estimates. Nearly equal large outputs can lose their small difference before division, and an extremely short interval may amplify that error. Removing a common large constant algebraically before evaluating an endpoint difference can preserve a change that floating-point arithmetic would otherwise lose. Verify such cases algebraically, retain exact fractions when working by hand, and do not treat extra displayed digits as extra measurement accuracy.
