8 topics. Each one opens with its key idea, then the detail you need to use it.
1. Interpreting the derivative in context
Key idea: Say what changes, when, how fast, in what units, and in which direction.
A derivative statement like $P'(5) = 2.4$ packs four facts into one line. Suppose $P(t)$ is a town's population in thousands and $t$ is years since 2020. Then $P'(5) = 2.4$ means that in 2025 the population is growing at a rate of 2.4 thousand people per year. Notice the pieces: the quantity (population), the moment ($t = 5$), the rate (2.4 thousand per year), and the direction (growing, because the value is positive). A negative value flips the direction word to falling or decreasing.
Units follow one rule. The derivative's units are the output's units divided by the input's units. If $D(t)$ is depth in feet and $t$ is hours, $D'$ is feet per hour and $D''$ is feet per hour per hour. Students often give the units of $D$ itself, or flip the fraction to hours per foot.
Keep instantaneous and average apart. $P'(5)$ describes the single instant $t = 5$. The change from $t = 0$ to $t = 5$ is $P(5) - P(0)$, and the average rate is that change divided by 5. Marginal cost is the business version of the same idea. If $C(x)$ is the cost of making $x$ items and $C'(300) = 7$, then the 301st item costs about 7 dollars extra. The word "about" matters, because the derivative is a linear estimate, not an exact cost.
2. Straight-line motion
Key idea: Matching signs for $v$ and $a$ mean speeding up; a sign change in $v$ means a change of direction.
A particle moving on a line has position $s(t)$, velocity $v(t) = s'(t)$, and acceleration $a(t) = v'(t) = s''(t)$. Velocity's sign gives direction: positive means moving right (or up), negative means moving left. Speed is $|v(t)|$, so speed is never negative even when velocity is.
Take $s(t) = t^3 - 9t^2 + 24t$ for $t \ge 0$. Then $v(t) = 3t^2 - 18t + 24 = 3(t - 2)(t - 4)$ and $a(t) = 6t - 18$. Velocity is positive before $t = 2$, negative between 2 and 4, and positive after 4. The particle changes direction at $t = 2$ and $t = 4$, because $v$ changes sign there. A zero of $v$ alone is not enough, since the sign has to switch.
Speeding up or slowing down depends on whether $v$ and $a$ agree in sign. At $t = 2.5$, $v = -2.25$ and $a = -3$. Both are negative, so the particle is speeding up even though it moves left. At $t = 3.5$, $v = -2.25$ and $a = 3$, so it is slowing down.
Total distance is not the same as displacement. From $t = 0$ to $t = 5$, the position goes from 0 to 20 at $t = 2$, back to 16 at $t = 4$, then up to 20 at $t = 5$. Displacement is $20 - 0 = 20$. Total distance adds each leg separately: $20 + 4 + 4 = 28$. Always split the trip at every turnaround.
3. Critical points, extreme values, and the First Derivative Test
Key idea: Critical points come from $f'$. A sign change in $f'$ makes a relative extremum, and a closed interval adds the endpoints as candidates.
A critical point of $f$ is an input $c$ in the domain where $f'(c) = 0$ or $f'(c)$ does not exist. You find them from the derivative, never from $f$ itself. Setting $f(x) = 0$ finds the $x$-intercepts, which is a different question.
Take $f(x) = x^4 - 4x^3$. Then $f'(x) = 4x^3 - 12x^2 = 4x^2(x - 3)$, so the critical points are $x = 0$ and $x = 3$. A critical point is only a candidate. The First Derivative Test decides what it is. If $f'$ changes from positive to negative at $c$, $f$ has a relative maximum there. If $f'$ changes from negative to positive, it is a relative minimum. If $f'$ keeps the same sign, there is no extremum. Here $4x^2$ is never negative, so the sign of $f'$ matches the sign of $x - 3$. That means $f'$ is negative on both sides of 0, so $x = 0$ is not an extremum. At $x = 3$, $f'$ goes from negative to positive, so $f$ has a relative minimum.
For absolute extrema on a closed interval, use the Extreme Value Theorem. A continuous function on $[a, b]$ must reach both an absolute maximum and an absolute minimum. They occur at critical points or endpoints, so test every candidate in $f$. On $[-1, 4]$: $f(-1) = 5$, $f(0) = 0$, $f(3) = -27$, $f(4) = 0$. The absolute maximum is 5 at $x = -1$, an endpoint, and the absolute minimum is $-27$ at $x = 3$.
4. Concavity and the Second Derivative Test
Key idea: The sign of $f''$ gives concavity, a sign change in $f''$ gives an inflection point, and $f''(c) = 0$ leaves the Second Derivative Test with no answer.
Concavity describes how the slope is changing. If $f''(x) > 0$, then $f'$ is increasing, so the slopes keep rising and the graph bends like a cup: concave up. If $f''(x) < 0$, the slopes are decreasing and the graph is concave down. Concavity is separate from direction. A graph with $f' > 0$ and $f'' < 0$ is still climbing, but it is leveling off, like the upper part of a hill.
An inflection point is where the concavity changes, which means $f''$ changes sign. Using $f(x) = x^4 - 4x^3$ again, $f''(x) = 12x^2 - 24x = 12x(x - 2)$. The sign of $f''$ is positive, then negative, then positive, so there are inflection points at $x = 0$ and $x = 2$. Be careful, because $f''(c) = 0$ alone does not prove an inflection point. For $g(x) = x^4$, $g''(0) = 0$, but $g''$ is positive on both sides, so the graph stays concave up.
The Second Derivative Test is a shortcut for classifying a critical point where $f'(c) = 0$. If $f''(c) > 0$, the graph is a cup there, so $f(c)$ is a relative minimum. If $f''(c) < 0$, it is a relative maximum. For our $f$, $f''(3) = 36 > 0$, which confirms the minimum at $x = 3$. But $f''(0) = 0$, and then the test says nothing. You must fall back on the First Derivative Test, which showed there is no extremum at 0.
5. Connecting f, f', and f'' (curve sketching)
Key idea: The height of $f'$ tells you where $f$ rises or falls. The slope of $f'$ tells you how $f$ bends.
Many questions hand you the graph of $f'$ and ask about $f$. The trap is reading it as if it were $f$. Translate each feature of the $f'$ graph one level down.
Where the graph of $f'$ is above the $x$-axis, $f$ is increasing. Where it is below, $f$ is decreasing. Where $f'$ crosses the axis from above to below, $f$ has a relative maximum. Where it crosses from below to above, $f$ has a relative minimum. If $f'$ only touches the axis and bounces back, $f$ has a horizontal tangent but no extremum.
Concavity comes from the slope of the $f'$ graph, not its height. Where $f'$ is rising, $f''$ is positive and $f$ is concave up. Where $f'$ is falling, $f$ is concave down. So the inflection points of $f$ sit at the peaks and valleys of the $f'$ graph, where $f'$ switches between rising and falling.
To sketch $f$ from a formula, build one table with intervals across the top and three rows: the sign of $f'$, the sign of $f''$, and the resulting shape. Each interval gets one of four shapes: increasing and concave up, increasing and concave down, decreasing and concave up, or decreasing and concave down. Then mark the extrema and inflection points and connect the pieces. For $f(x) = x^4 - 4x^3$, the table reads decreasing and concave up before 0, decreasing and concave down from 0 to 2, decreasing and concave up from 2 to 3, and increasing and concave up after 3.
6. Related rates
Key idea: Write the equation, differentiate with respect to $t$, and only then plug in the numbers for the instant.
In related rates, two or more quantities change over time, and an equation ties them together. You know one rate and need another at a specific instant.
First, draw a picture and name the variables that change. Second, write an equation linking them that is true at every moment, not just the instant in question. Third, differentiate both sides with respect to $t$, using the chain rule on every variable. Fourth, substitute the values for the instant and solve.
Suppose a spherical balloon is inflated at 100 cubic centimeters per second. How fast is the radius growing when $r = 5$ cm? The link is $V = \frac{4}{3}\pi r^3$. Differentiating gives $\frac{dV}{dt} = 4\pi r^2 \frac{dr}{dt}$. Now substitute: $100 = 4\pi(25)\frac{dr}{dt}$, so $\frac{dr}{dt} = \frac{1}{\pi} \approx 0.318$ cm per second.
The most common error is plugging in $r = 5$ before differentiating. That turns the radius into a constant, and its derivative becomes zero. Only true constants, like a ladder's fixed length, can go in early. Two other errors show up in ladder problems. Students assume both ends move at the same rate, but in $x^2 + y^2 = L^2$ the rates are linked by the ratio $\frac{x}{y}$, not equal. Students also flip that ratio. Solving carefully gives $\frac{dy}{dt} = -\frac{x}{y}\frac{dx}{dt}$, and the negative sign tells you the top is sliding down.
7. Optimization
Key idea: Use the constraint to get one variable, set the domain, find and justify the extremum, and answer the exact question asked.
Optimization asks for the largest or smallest value of a quantity under a limit, such as a fixed amount of fencing or material. The hard part is building the function.
Work in this order. Name the quantity to optimize and write a formula for it. That formula usually has two variables. Write the constraint equation, solve it for one variable, and substitute so the function has a single variable. State the domain that makes physical sense. Then find critical points, test them, and answer the question that was actually asked.
Suppose you have 120 meters of fencing for a rectangular pen along a river, with no fence on the river side. Let $x$ be each side perpendicular to the river and $y$ the side parallel to it. The constraint is $2x + y = 120$, so $y = 120 - 2x$. The area is $A(x) = x(120 - 2x) = 120x - 2x^2$ on $0 \le x \le 60$. Then $A'(x) = 120 - 4x = 0$ gives $x = 30$, so $y = 60$ and $A = 1800$ square meters. Since $A''(x) = -4 < 0$, this critical point is the maximum.
Two habits cost points. One is using the wrong constraint, such as fencing all four sides. The other is guessing equal sides. Three sides of 40 meters each give only 1600 square meters, which is less than 1800. On a free-response question, you also need a reason the value is a maximum, such as the sign change of $A'$ or the candidates test.
8. L'Hôpital's Rule
Key idea: Confirm $\frac{0}{0}$ or $\frac{\infty}{\infty}$ first, differentiate top and bottom separately, and rewrite other forms as a fraction before you start.
L'Hôpital's Rule evaluates limits that come out as $\frac{0}{0}$ or $\frac{\infty}{\infty}$ when you substitute. In those cases, $\lim \frac{f(x)}{g(x)} = \lim \frac{f'(x)}{g'(x)}$, provided the new limit exists. The rule also needs $f$ and $g$ to be differentiable near the point, and $g'(x)$ must not be 0 there. You differentiate the top and the bottom separately. Do not use the quotient rule, which answers a different question.
Always check the form first. If substitution gives a number like $\frac{3}{2}$ or $\frac{5}{0}$, the rule does not apply, and using it gives a wrong answer. A $\frac{0}{0}$ form also does not mean the limit is 0 or infinite. The top and bottom are both shrinking toward zero, and the derivatives compare how fast.
You may apply the rule more than once. For $\lim_{x \to 0} \frac{1 - \cos x}{x^2}$, substitution gives $\frac{0}{0}$. One round gives $\frac{\sin x}{2x}$, still $\frac{0}{0}$. A second round gives $\frac{\cos x}{2}$, which goes to $\frac{1}{2}$.
BC questions also use other indeterminate forms, which you must rewrite as a fraction first. For $0 \cdot \infty$, as in $\lim_{x \to 0^+} x \ln x$, write it as $\frac{\ln x}{1/x}$, which has the form $\frac{-\infty}{\infty}$. Differentiating gives $\frac{1/x}{-1/x^2} = -x$, which goes to 0. For a $1^\infty$ form, take the natural log, find that limit, then raise $e$ to it. On a free-response question, write the limits of the top and bottom separately to show the form.