Intuition The ONE core ideaEvery measurement is really a little interval — a best guess with a "give or take" attached — and error analysis is just the bookkeeping that tracks the width of that interval. Before you can do that bookkeeping, you must be fluent in a handful of symbols (sums, means, absolute value, fractions), so this page earns each one from scratch.
This is the toolbox page for the parent topic on errors . The parent throws around ∑ ∣ ∣
Before any symbol, hold this picture in your head: numbers live on a straight ruled line, and a measurement is a dot on that line. The true value is another dot we can't see. "Error" is simply the distance between two dots .
Definition Value, true value, measured value
A value is just a number with a unit, e.g. 5.2
The true value is the exact answer nature holds — an invisible dot we chase but never quite touch.
The measured value is the dot our instrument actually lands on.
The picture: two dots on a line. The gap between them is what the whole topic is about.
The parent writes a 1 , a 2 , … , a n
Definition Subscript notation
A ==letter like a without committing to specific numbers.
The ==subscript i label , a seat number. a 1 a 2 not a multiplication or a power — it just names which reading.
The letter ==n how many readings you took in total. If you timed a pendulum 5 n = 5 a 1 a 5
Intuition Why we need labels at allIf you measure something once, you get one number and there's nothing to average. Errors only become tameable when you repeat — and to talk about "each repeat" you need a way to point at reading number i n
Question: In a 3 3 It is a label naming the third reading — not "a a 3
The very first formula in the parent is a mean = n 1 ∑ a i ∑ S um) is the star of this show.
Intuition Why a special symbol instead of just "+ ⋯ + n … ∑ precisely where to start, where to stop, and what to add — no ambiguity.
Worked example Warm-up sumWith readings 2.63 , 2.56 , 2.42 , 2.71 , 2.80 ∑ i = 1 5 a i = 2.63 + 2.56 + 2.42 + 2.71 + 2.80 = 13.12 5 13.12
Definition Arithmetic mean (average)a mean = n 1 ∑ i = 1 n a i = n a 1 + a 2 + ⋯ + a n WHAT: add up all readings, then split the total equally among the n The picture: if each reading were a coin stacked on the number line, the mean is the point where the stacks perfectly balance — the see-saw pivot.
n 13.12 typical reading. Dividing by n
Recall Check the balance ideaFor 2.63 , 2.56 , 2.42 , 2.71 , 2.80 13.12/5 = ? Reveal: 2.624 2.80 2.71 mean you'll meet again when we quantify scatter.
The parent writes Δ a i = ∣ a mean − a i ∣ Δ
Definition Absolute valueThe bars ∣ x ∣ x x ∣ + 0.086 ∣ = 0.086 , ∣ − 0.064 ∣ = 0.064 The picture: fold the number line at 0 distance from 0
Intuition Why we throw away the sign hereWe want to know how far a reading missed the mean, not which side it missed on. A reading 0.06 0.06
∣ a − b ∣ a b Why it feels right: you compute a − b The fix: if that comes out negative, flip its sign. ∣2.42 − 2.624∣ = ∣ − 0.204 ∣ = 0.204 − 0.204
Δ Δ D for D ifference) attached to a quantity means "the small change or gap in that quantity ." So Δ a a a The picture: a tiny bracket ⊢ ⊣
Intuition Why one dedicated symbol for "the wobble"Errors are about gaps and small amounts. Instead of writing "the uncertainty in a Δ Δ Δ how errors combine through calculations .
Question: What does Δ a mean The typical size of the wobble — how far a reading usually strays from the best guess.
The parent's relative error is δ r = a mean Δ a mean δ idea of a ratio.
Definition Ratio / fraction as "per unit"A fraction q p p for each single unit of q how big is the error compared to the thing itself?
2.62 s 0.11 s = 0.042 cancel — top and bottom carry the same unit — so the answer is a bare number. That "no units left" is exactly why relative error lets you compare a length measurement with a mass measurement.
1 1 2 half the wire — a disaster. The same 1 2 1
Definition Lower-case deltaδ δ relative/fractional kind of error (a plain ratio), to visually distinguish it from the Δ δ r δ %
Definition Per-cent = "per hundred"The word percent literally means "out of 100 100 % 100 1 0.042 × 100 = 4.2 , written 4.2% The picture: the same slice of a pie, but the pie is now cut into 100 1
Mnemonic A Rabbit PercentA bsolute (has units) → R elative (divide, units cancel) → P ercent (×100). Each step makes the number more comparable than the last.
Relative error small delta
Read top to bottom: the number line and subscripts are bedrock; they build the sum and the mean; the mean plus absolute value plus Δ 100 is the errors topic.
Cover the right side and see if you can answer each before revealing.
∑ i = 1 n a i Add up readings a 1 a n 1 n
What does the subscript in a 3 The third reading — it is a seat number, not a power or a multiplication.
How do you compute the arithmetic mean of n Add them all with ∑ n
Why divide the sum by n To strip out "how many readings" and get the size of a typical one — the balance point.
What does ∣ − 0.2 ∣ 0.2
Why use absolute value when finding how far a reading missed? We want the size of the miss; keeping signs would let plus and minus misses cancel and hide the error.
What does the symbol Δ A small change or uncertainty — the size of the wobble, not a value.
Why is relative error dimensionless? It divides an error by a value with the same unit, so the units cancel, leaving a pure number.
Turn a relative error of 0.042 Multiply by 100 4.2%
Which symbol (Δ δ Δ δ
