Intuition The one core idea
To average infinitely many function values, we add them all up with an integral (a "continuous sum") and divide by the length of the interval. So this whole topic is built from just two ideas glued together: summing (which becomes the integral) and dividing by how many (which becomes dividing by the width b − a ).
This page assumes nothing . On the parent note you will meet a formula that averages a function; the shorthand for "the average value of f " is written f avg — we define this symbol fully in §8, so treat it for now just as a name for the answer we are building toward . Before you meet that formula on the parent note , we build every single symbol in it, one at a time, each with a picture and a reason it has to be there.
A function is a machine: you feed it a number x , it gives back exactly one number, written f ( x ) (read "f of x"). The letter f is the machine's name; x is the input; f ( x ) is the output.
Picture it as a curve. Every point on the horizontal axis is an input x ; the height of the curve above that point is the output f ( x ) .
Intuition Why the topic needs this
"Average value" is the average of all those heights . If you can't see f ( x ) as a height, you can't picture averaging it. Everything downstream is about the heights of this one curve.
Why so-many-values matters A function has one height for every x in the interval — that's infinitely many numbers, which is exactly the problem average value solves.
Definition Closed interval
[ a , b ]
[ a , b ] means "all numbers from a up to b , including both ends." Here a is the left endpoint (start) and b is the right endpoint (end). We always take b > a , so the interval runs left-to-right.
On the picture: a and b are two marks on the horizontal axis. The stretch of curve between them is the only part we care about.
Intuition Why the topic needs this
You can only average over a definite stretch. "The average of x 2 " is a meaningless question — "the average of x 2 on [ 0 , 3 ] " has one answer. The interval is what makes the question well-posed.
Definition Width of the interval
b − a is the length of the interval — how far you travel from the start a to the end b . Because b > a , this number is always positive .
Picture: the length of the base of the shaded region, measured along the horizontal axis.
Intuition Why the topic needs this
When you average n ordinary numbers you divide by n ("how many"). For a continuous function the "how many" becomes the width b − a . It is the denominator of the whole formula — miss it and you get an area, not an average.
a − b instead of b − a
Why it tempts: under pressure the order of letters gets scrambled.
Why it's wrong: a width can't be negative. a − b would be negative when b > a .
Fix: big minus small. Width = b − a > 0 .
Now the biggest symbol. Let's build it in pieces.
Intuition Why "area" answers "sum"
Imagine slicing the shaded region into thin vertical strips. Each strip is (almost) a rectangle: its height is f ( x ) and its width is a tiny number we call d x . Its area is height × width = f ( x ) d x . Add up every strip's area and you have the total area. That "add up all the strips" is what the stretched-S symbol ∫ means — it literally started life as a long "S" for Sum .
Definition The definite integral, symbol by symbol
∫ a b f ( x ) d x
∫ — "add up all the strips" (a continuous sum).
a (bottom) and b (top) — start and stop adding here.
f ( x ) — the height of each strip.
d x — the infinitely thin width of each strip.
Together: the total signed area between the curve and the horizontal axis, from a to b .
"Signed" matters: strips below the axis (where f ( x ) is negative) count as negative area. See §6.
Intuition Why the topic needs this
The integral is the "add up all the (infinitely many) heights" step. Average value = this total, spread evenly. Without the integral there is no way to sum infinitely many heights — see Definite Integral as Area and Riemann Sums for how the sum is made rigorous.
The parent's derivation passes through this, so let's earn every symbol.
Definition Sigma notation
∑
i = 1 ∑ n f ( x i ) means "add up f ( x 1 ) + f ( x 2 ) + ⋯ + f ( x n ) ." The ∑ is a capital Greek "S" (again for Sum ); i counts from 1 to n ; x i is the i -th sample point.
Definition Where does each sample point
x i sit?
Chop [ a , b ] into n strips, each of width Δ x = n b − a . Inside the i -th strip you pick one input to measure the height at — call it x i . You are free to pick:
the left endpoint of the strip: x i = a + ( i − 1 ) Δ x ,
the right endpoint : x i = a + i Δ x ,
the midpoint , or literally any point inside the strip.
f ( x i ) is then the height of that one chosen sample, used for the whole strip.
Intuition Why the choice stops mattering in the limit
Different choices give slightly different staircases (look at the figure: left-endpoint bars vs right-endpoint bars). But as the strips get thinner (n → ∞ ), every choice inside a strip is squeezed to the same place — so all choices converge to the same integral . That is exactly why the notation ∑ f ( x i ) Δ x is allowed to be a little vague about x i : the answer doesn't depend on it.
Δ x and n
n = how many strips we chop the interval into (a finite whole number).
Δ x (read "delta x") = the width of each strip = n b − a . "Δ " means "a change in / a chunk of."
Definition When is this limit guaranteed to exist? (integrability)
The staircase sum only settles to a single number if the function is well-behaved . The safe, sufficient condition we use throughout this topic is:
f is == continuous == on [ a , b ] .
"Continuous" means you can draw the curve without lifting your pen — no jumps, no holes, no spikes to infinity. A continuous function on a closed interval is always integrable , so its Riemann sums have a limit and ∫ a b f d x makes sense.
every function has an average value
Why it tempts: the formula looks like it always works.
Why it's wrong: if f has a jump or blows up to infinity, the Riemann sums may not settle to any number, so the integral — and the average — can fail to exist.
Fix: check f is continuous (or at least integrable) on [ a , b ] first. Every example in this topic is continuous, so we're safe.
Intuition Why the topic needs this
This is the bridge in the derivation. The finite average n 1 ∑ f ( x i ) (which we already trust) turns into b − a 1 ∑ f ( x i ) Δ x , and the sum turns into ∫ . Every step in the parent's "HOW we derive" section lives here.
You must handle every case, not just curves that stay above the axis.
Definition Sign of a height
f ( x ) > 0 : strip is above the axis → area counts positive .
f ( x ) < 0 : strip is below the axis → area counts negative .
f ( x ) = 0 : strip has zero height → contributes nothing.
Intuition Why the topic needs this
The average value can be negative (if the curve mostly sits below the axis) or exactly zero (if positive and negative parts cancel). Example: f ( x ) = sin x on [ 0 , 2 π ] has average 0 — the hump above cancels the dip below. If you only ever pictured positive curves, that answer would look impossible.
Before we compute that example, we need one more tool — a way to actually evaluate an integral without summing infinitely many strips by hand.
Definition Antiderivative and the evaluation bracket
An antiderivative of f is any function F whose slope is f everywhere — that is, F is a function that "undoes" the rate of change back to f . For example, − cos x is an antiderivative of sin x , because the slope of − cos x is sin x .
The Fundamental Theorem of Calculus then says: the total signed area equals the antiderivative measured at b minus at a . We write this with the evaluation bracket :
∫ a b f ( x ) d x = [ F ( x ) ] a b = F ( b ) − F ( a ) .
The tall bracket [ ⋯ ] a b is just shorthand for "plug in the top number, subtract plug-in the bottom number." This is the shortcut that turns the infinite Riemann sum into simple arithmetic.
Worked example A zero-average case
Using F ( x ) = − cos x (an antiderivative of sin x ):
∫ 0 2 π sin x d x = [ − cos x ] 0 2 π = ( − cos 2 π ) − ( − cos 0 ) = ( − 1 ) − ( − 1 ) = 0.
So f avg = 2 π 1 ⋅ 0 = 0 . The positive hump and negative dip exactly cancel. ✓
n → ∞ lim ( expression ) asks: "as n grows without bound, what single number does the expression settle down to?" It's the value you get closer and closer to, never needing to reach infinity itself.
Picture: watch the staircase of rectangles from §5. With n = 4 it's crude; n = 20 better; n = 1000 you can't tell it from the curve. The limit is that "perfect fit."
Intuition Why the topic needs this
"Average of infinitely many" is not something we can compute directly — we compute the average of n samples and take the limit . The limit is what upgrades an approximation into an exact continuous average.
Definition Notation for "average"
A bar over a letter, like y ˉ or f ˉ , is standard shorthand for "the average of." The parent also writes it as f avg . Same thing, clearer name. This is the symbol we flagged in the opening line — now fully earned.
Now every piece of
f avg = b − a 1 ∫ a b f ( x ) d x
is defined: f avg (§8) is the average height; b − a 1 (§3) spreads the total over the width; ∫ a b (§4, evaluated via the bracket in §6) sums all the heights; f ( x ) d x (§1, §4) is one strip's contribution; and the whole thing is trustworthy because f is continuous (§5). Area over width — nothing unexplained remains.
The foundations stack up in this order. Read it as "the box below is needed before the box above works":
Function as a height f ( x ) (§1) and the interval [ a , b ] (§2) are the raw material.
From the interval you get the width b − a (§3) — the denominator.
Chopping the interval and summing sampled heights gives the Riemann sum ∑ f ( x i ) Δ x (§5), which needs a sample-point choice and continuity/integrability to be well-defined.
Taking the limit n → ∞ (§7) turns the Riemann sum into the definite integral ∫ a b f d x (§4), evaluated with the antiderivative bracket (§6).
Integral over width gives the average value f avg (§8), which in turn powers the Mean Value Theorem for Integrals .
Recall Self-test: can you answer each before peeking?
What does f ( x ) mean in plain words, and what does it look like? The output of the machine f fed input x ; it is the height of the curve above the point x .
What does the interval [ a , b ] include? Every number from a to b , including both endpoints, with b > a .
Why must the width be b − a and never a − b ? A width is positive; since b > a , big-minus-small = b − a > 0 .
In ∫ a b f ( x ) d x , what is each of ∫ , a , b , f ( x ) , d x ? ∫ = sum all strips; a , b = start/stop; f ( x ) = strip height; d x = strip width.
Why does the integral compute a "sum"? Each thin strip has area f ( x ) d x ; the integral adds every strip → total area = continuous sum of heights.
Where inside each strip does the sample point x i have to sit? Anywhere in the strip — left endpoint, right endpoint, midpoint, or any point; in the limit the choice doesn't matter.
What condition on f guarantees the Riemann-sum limit (the integral) exists? f is continuous on [ a , b ] (a sufficient condition for integrability).
What does ∑ i = 1 n f ( x i ) Δ x become as n → ∞ ? The definite integral ∫ a b f ( x ) d x — the Riemann sum limit.
What is Δ x in terms of n ? Δ x = n b − a , the width of each of the n strips.
What does the evaluation bracket [ F ( x ) ] a b mean? F ( b ) − F ( a ) : plug in the top, subtract plug in the bottom (an antiderivative F has slope f ).
What does lim n → ∞ ask you to find? The single number the expression settles toward as n grows without bound.
What happens to area where f ( x ) < 0 ? It counts as negative area, which can drag the average below zero or cancel positive parts to zero.
What is the average value of sin x on [ 0 , 2 π ] , and why? 0 — the hump above the axis cancels the dip below.