Squeeze theorem for sequences: Visual walkthrough

Before any symbol, one plain-language promise:

Step 0 — What is a "sequence" and what is a "limit"?

WHAT. A sequence is just an endless ordered list of numbers, one for each counting number $n = 1, 2, 3, \dots$ . We write the $n$ -th number as $b_{n}$ (read " $b$ sub $n$ "). The whole list is $(b_{n})$ .

WHY start here. The parent note assumed you know what $b_{n} \to L$ means. We refuse to assume. Every later step leans on this one definition, so we anchor it to a picture first.

PICTURE. Plot each $b_{n}$ as a dot at height $b_{n}$ above its index $n$ . A limit $L$ is a horizontal line the dots eventually hug. "Eventually hug" is made precise by an $ε$ -band (epsilon-band): pick any tiny height $ε > 0$ , draw the strip from $L - ε$ to $L + ε$ ; from some index onward, every dot lands inside that strip.

This is the full formal machinery — see Limit of a sequence (epsilon-N definition). Everything below is just this picture, applied three times.

Step 1 — Draw the three sequences and the trap

WHAT. Introduce the three lists by name:

$a_{n}$ — the lower wall (bottom bread of the sandwich),
$b_{n}$ — the middle sequence, the one we actually care about (the filling),
$c_{n}$ — the upper wall (top bread).

WHY these three. The middle list $b_{n}$ (imagine $\frac{s i n n}{n}$ ) may wiggle too wildly to follow directly. So we don't chase it — we cage it between two calmer lists we understand.

PICTURE. At every index $n$ the three dots stack in order: $a_{n}$ at the bottom, $b_{n}$ squeezed in the gap, $c_{n}$ on top. That vertical ordering is the trap.

Step 2 — Demand that both walls land on the same number $L$

WHAT. Require $a_{n} \to L$ and $c_{n} \to L$ — the same $L$ . Draw one horizontal line at height $L$ ; both wall-sequences must funnel onto it.

WHY the same $L$ . If the walls stop at different heights (say $a_{n} \to 0$ , $c_{n} \to 5$ ), the gap between them never closes, and $b_{n}$ can loiter anywhere in that gap. No conclusion possible. The shared target is what makes the gap shrink to nothing.

PICTURE. Two funnels — the lower wall rising, the upper wall falling — both aiming at the single line $L$ . The shaded region between them narrows toward a point.

Step 3 — Pick a challenge strip $ε$ and pin the lower wall

WHAT. Fix any $ε > 0$ (think: an opponent dares you with a strip). Because $a_{n} \to L$ , from some index $N_{1}$ onward every $a_{n}$ sits inside the strip $(L - ε, L + ε)$ .

WHY only keep the left half. The lower wall's job is to hold $b_{n}$ up from below. We only need the fact that $a_{n}$ can't drop below $L - ε$ . So we extract just the left inequality and discard the rest.

PICTURE. Highlight the floor line $L - ε$ . Past $N_{1}$ , every red lower-wall dot stays above it.

Step 4 — Pin the upper wall with the same $ε$

WHAT. Same move, mirror image. Because $c_{n} \to L$ , there is $N_{2}$ so that past it every $c_{n}$ lies inside $(L - ε, L + ε)$ . Keep only the right half: $c_{n} < L + ε$ .

WHY the right half now. The upper wall's job is to hold $b_{n}$ down from above. All we need is that $c_{n}$ can't climb past $L + ε$ .

PICTURE. Highlight the ceiling line $L + ε$ . Past $N_{2}$ , every violet upper-wall dot stays below it.

Step 5 — Stack all three facts at once

WHAT. We now have three separate promises, each valid from its own starting index:

the trap ( $a_{n} \leq b_{n} \leq c_{n}$ ) from $N_{0}$ ,
the floor ( $L - ε < a_{n}$ ) from $N_{1}$ ,
the ceiling ( $c_{n} < L + ε$ ) from $N_{2}$ .

To use them together, wait until all three are active. Let $N = max (N_{0}, N_{1}, N_{2}),$ the latest of the three starting indices. From $N$ onward, all three hold simultaneously.

WHY the max. A chain is only as ready as its slowest link. Taking the largest index guarantees no promise has "not started yet."

PICTURE. One combined chain of inequalities, read left to right along the number line at a single index $n \geq N$ :

Step 6 — Read off the outermost pieces

WHAT. In that chain, ignore the busy middle and read only the two ends touching $b_{n}$ : $L - ε < b_{n} < L + ε .$

WHY this is the finish. This says $b_{n}$ sits inside the very strip the opponent challenged us with — i.e. $b_{n}$ 's distance from $L$ is under $ε$ .

PICTURE. The middle dot $b_{n}$ is now sandwiched inside the $ε$ -band. It had no room to escape: floor below, ceiling above.

Step 7 — Conclude (and note what we secretly proved for free)

WHAT. The opponent's $ε$ was arbitrary — we never used a specific value. So for every $ε > 0$ we can produce the index $N$ that forces $∣ b_{n} - L ∣ < ε$ . That is the Step 0 definition. Hence $b_{n}$ converges and its limit is $L$ . $■$

WHY "converges" is a bonus. We didn't assume $b_{n}$ had a limit and then find it — we proved existence of the limit and its value in one stroke. That's why squeeze is so powerful for wild sequences like $\frac{s i n n}{n}$ : it hands you convergence and the value together.

Recall Where the

$ε$ -choices came from (self-test) Which fact did the lower wall supply? ::: $L - ε < a_{n}$ for $n \geq N_{1}$ . Which fact did the upper wall supply? ::: $c_{n} < L + ε$ for $n \geq N_{2}$ . Why take $N = max (N_{0}, N_{1}, N_{2})$ ? ::: So all three promises are active at once. What made the conclusion "for every $ε$ " legal? ::: We fixed $ε$ arbitrarily and never used its size.

Edge & degenerate cases (each earns its own look)

The one-picture summary

Everything above, compressed: two funnels closing on the line $L$ , the middle dots crushed into the shrinking $ε$ -band, with $N$ marked where the crush becomes permanent.

Recall Feynman retelling — the walkthrough in plain words

Someone challenges you: "I bet the middle list ever escapes this thin strip around $L$ ." You answer in four moves. First (Steps 3–4), you note the bottom wall eventually rises above the strip's floor, and the top wall eventually drops below the strip's ceiling — because both walls are heading exactly to $L$ . Second (Step 5), you wait long enough (index $N$ ) that both of those are true at the same time, plus the middle is genuinely between the walls. Third (Step 6), you look at the middle list: it's above the floor (it sits above the bottom wall) and below the ceiling (it sits below the top wall) — so it's inside the strip. Fourth (Step 7), since you never cared how thin the strip was, the middle list is trapped inside any strip you're handed. That is the definition of heading to $L$ . The middle never had a choice.

Connections

Parent: Squeeze theorem for sequences — the statement, examples, and pitfalls this page visualises.
Limit of a sequence (epsilon-N definition) — Step 0 is this definition; the whole proof is pure $ε$ – $N$ .
Bounded sequences — where the walls come from ( $∣ sin n ∣ \leq 1$ , etc.).
Algebra of limits (sum, product, quotient) — the gentler alternative when the middle sequence is tame.
Squeeze theorem for functions — the same picture with a continuous variable instead of $n$ .
Standard limits ( n-th roots, n!/n^n, ln n / n ) — the classic sequences these walls tame.
Monotone convergence theorem — a different existence tool for when no walls are handy.