📊 Section 1 -- Interactive Simulation
Controls
Function
Display
💡 Section 2 -- The Idea, Step by Step
One idea, climbing gently from a road-trip speedometer to the formal theorem.
Imagine a road trip: you cover 180 km in exactly 2 hours, so your average speed was 90 km/h. Along the way you sped up, slowed down, maybe stopped for coffee, so the needle was rarely sitting on 90. But here is the catch a highway speed camera quietly relies on: at some single instant you must have been going exactly 90 km/h. You cannot average 90 over the whole trip without actually touching 90 somewhere in the middle. That inescapable "somewhere" is the entire idea of the Mean Value Theorem.
Now translate "average speed" into the language of a graph. Plot a quantity $f$ against the input (say distance against time). The starting point $(a,f(a))$ and the finishing point $(b,f(b))$ are joined by a straight secant line, and its slope is the average rate of change over the interval:
The steepness of the curve at one single instant is instead the slope of the tangent line there, written $f'(c)$. The Mean Value Theorem promises that at least one interior point $c$ exists where the tangent is exactly parallel to the secant — where the instantaneous rate equals the average rate. Try real numbers: for $f(x)=x^2$ on $[1,3]$ the average rate is $\frac{9-1}{3-1}=4$, and solving $f'(c)=2c=4$ gives $c=2$, comfortably inside the interval.
Stated precisely: if $f$ is continuous on the closed interval $[a,b]$ and differentiable on the open interval $(a,b)$, then there is some $c\in(a,b)$ with $f'(c)=\dfrac{f(b)-f(a)}{b-a}$. Both conditions earn their keep — a sharp corner like $|x|$ or a jump in the graph breaks the guarantee. Rolle's Theorem is just the special case where the two endpoints sit at the same height, $f(a)=f(b)$: the average rate is then $0$, so some tangent must be perfectly horizontal, $f'(c)=0$. In the sim above, the $a$ and $b$ sliders set the two endpoints, the gold dashed line is the secant, and the purple line marks the tangent at the $c$ the page solves for.
Try this in the sim above. (1) Choose $\sin(x)$ and set $a=0$, $b\approx3.14$ so both endpoints sit at zero; the gold secant goes flat and the purple tangent lands horizontal near $c=\pi/2$ — that is Rolle's Theorem in action. (2) Press ▶ Animate to sweep the right endpoint $b$ and watch $c$ glide along the curve in real time, always keeping its tangent parallel to the secant. (3) Switch to $x^2-2x$ and drag $a$ and $b$ around: the $c$ readout always lands exactly at the midpoint $(a+b)/2$ — a quirk special to parabolas that is never true in general.
📐 Section 3 -- Mean Value Theorem
If $f$ is continuous on $[a,b]$ and differentiable on $(a,b)$, then there exists $c\in(a,b)$ such that $$f'(c)=\frac{f(b)-f(a)}{b-a}.$$
Rolle's Theorem (special case): if additionally $f(a)=f(b)$, then $\exists\,c\in(a,b)$ with $f'(c)=0$.
| Symbol | Meaning | Condition |
|---|---|---|
| $f$ cts on $[a,b]$ | No jumps, holes, or vertical asymptotes | Closed interval |
| $f$ diff on $(a,b)$ | No cusps or corners inside | Open interval (strict) |
| $c\in(a,b)$ | Guaranteed interior point | Existence, not uniqueness |
| $f'(c)$ | Instantaneous rate at $c$ | Equals average rate |
The secant line through $(a,f(a))$ and $(b,f(b))$ has slope $(f(b)-f(a))/(b-a)$. MVT says: somewhere between $a$ and $b$, the tangent line is parallel to this secant. Intuitively: if your average speed was 80 km/h, at some instant you were traveling exactly 80 km/h.
Define $g(x)=f(x)-[f(a)+(f(b)-f(a))/(b-a)\cdot(x-a)]$ (subtract the secant line). Then $g(a)=0$ and $g(b)=0$. $g$ satisfies Rolle's hypotheses. By Rolle: $\exists c$ with $g'(c)=0$, i.e., $f'(c)=(f(b)-f(a))/(b-a)$.
By extreme value theorem, $f$ attains its max $M$ and min $m$ on $[a,b]$. If $M=m$, $f$ is constant and $f'\equiv0$. If $M>m$, at least one of them is attained at interior $c\in(a,b)$. At a local extremum in the interior, $f'(c)=0$ (Fermat's theorem). $\square$
$f'(x)>0$ on $(a,b)\Rightarrow f$ strictly increasing. $f'(x)=0$ on $(a,b)\Rightarrow f$ constant. $f'(x)=g'(x)$ on $(a,b)\Rightarrow f-g=$ const. These all follow from MVT applied to the appropriate function.
For two differentiable functions $f,g$ on $[a,b]$ with $g'\neq0$: $\exists c\in(a,b)$ with $f'(c)/g'(c)=(f(b)-f(a))/(g(b)-g(a))$. This is used to prove L'Hopital's rule.
$f(x)=x^3-x$ on $[-1,2]$. $f(-1)=0$, $f(2)=6$. Average slope $=(6-0)/(2-(-1))=2$. Find $c$: $f'(c)=3c^2-1=2\Rightarrow c^2=1\Rightarrow c=\pm1$. Only $c=1\in(-1,2)$ ✓. The tangent at $x=1$ has slope 2, parallel to the secant.
❓ Section 4 -- FAQ
No -- MVT guarantees existence of at least one such $c$, but there may be many. For $f(x)=\sin x$ on $[0,2\pi]$, the secant slope is 0 and $f'(c)=\cos c=0$ at $c=\pi/2$ and $c=3\pi/2$ -- two values. The theorem is purely existential.
Key takeaway: MVT guarantees existence of at least one c, possibly many. Not uniqueness.The gold secant line connects $(a,f(a))$ to $(b,f(b))$. The purple tangent line at $c$ is parallel to the secant. As you move the sliders for $a$ and $b$, $c$ updates in real time (found numerically). The animation sweeps $c$ from $a$ to $b$ showing the tangent rotating to match the secant slope.
Key takeaway: Secant slope = f'(c). Tangent at c is parallel to secant. Move a and b to see c update live.Speed cameras: if your average speed between two points exceeded the limit, MVT proves you were speeding at some moment. Numerical analysis: error bounds for numerical differentiation. Physics: MVT justifies $\Delta x = v(t^*)\Delta t$ for some $t^*$ in each time step. Medicine: MVT used in pharmacokinetic modeling (concentration rates).
Key takeaway: Speed cameras, error bounds, pharmacokinetics, numerical analysis -- MVT justifies them all.$f(x)=1/x$ on $(0,1)$ is differentiable but not continuous on $[0,1]$ (blow-up at 0). $f(0.1)=10$, $f(0.9)\approx1.11$. Average slope $\approx-9.87/0.8$. But $f'(c)=-1/c^2$ and you can check the MVT equation has no solution in $(0,1)$. Continuity on the closed interval is essential.
Key takeaway: MVT requires continuity on CLOSED interval [a,b]. Fails if f blows up or has jump at endpoint.$f(1)=1$, $f(3)=9$. Average slope $=(9-1)/(3-1)=4$. Find $c$: $f'(c)=2c=4\Rightarrow c=2\in(1,3)$ ✓. Unique here. Tangent at $x=2$: slope 4, same as secant through $(1,1)$ and $(3,9)$. Check: secant equation $y=4x-3$; tangent $y=4x-4$. Parallel (same slope) ✓.
Key takeaway: f(x)=x^2 on [1,3]: average slope=4, f'(c)=2c=4, so c=2. Unique MVT point for this quadratic.MVT is the bridge: if $F'=f$ on $(a,b)$ and $f$ is continuous, then $F(b)-F(a)=f(c)(b-a)$ for some $c$ by MVT. Taking refinements as the partition gets finer leads to $F(b)-F(a)=\int_a^b f(x)\,dx$. Also: MVT in higher dimensions becomes the Mean Value Inequality $\|f(b)-f(a)\|\leq\|Df\|\cdot\|b-a\|$.
Key takeaway: MVT is the key lemma in proving FTC. In higher dimensions: ||f(b)-f(a)|| <= sup||Df|| * ||b-a||.