Squeeze Theorem
A function too wild to evaluate directly can still be pinned down, if it is
trapped between two functions that agree in the limit. Zoom in and watch the
trap close.
x^2 sin(1/x)
x sin(1/x)
x^2 cos(5/x)
sin(1/x) — no squeeze
the trap, between f and h
g , trapped
the bounds
The trap closing
Bounds at x = δ , and the gap between them
δ lower f
upper h gap
Squeeze Theorem
Three functions. If everywhere near
a , and the two outer ones have the same limit L at
a , then g has nowhere else to go: its limit is L too.
The shaded region is the trap. Drag δ down and watch it close.
What to watch for
The wild function is never evaluated at a .
is undefined at 0 — sin(1/0) means
nothing. The theorem never asks for it. It concludes from the
neighbours.
The oscillation does not stop; the room for it does. Tick
"Rescale the vertical axis" and g wobbles just as violently at
δ = 0.001 as at δ = 1. Untick it and the wobbling is
crushed flat. Both pictures are true, and students who have only seen
the second think the function calms down. It does not.
The hypothesis has to hold. Try the sin(1/x) preset: it is
genuinely trapped between −1 and 1, but those bounds never meet, so the
gap stays 2 and the theorem concludes nothing. Bounds are not enough —
they must agree in the limit .
Putting this in a slide
?g=x%5E2+sin(1%2Fx)&lower=-x%5E2&upper=x%5E2&a=0
&embed=true&controls=delta,table
Note that a plus sign inside a function must be written %2B,
and a slash %2F. Every parameter is in
docs/url-api.md.
Keyboard
← → Change δ (focus the slider first)
Space Run or stop the zoom
? This dialog
Esc Close
Close