This page is the drill floor . The parent note built the rule and its proof; here we hit every kind of problem the rule can face, so no exam surprises you. If a symbol appears you have not met, the parent note or Indeterminate Forms defines it.
Intuition What a "scenario" means here
A limit is a shape of trouble . 0 0 is one shape, ∞ − ∞ another, 1 ∞ another. Each shape has its own first move before L'Hôpital can even be used. The matrix below lists every shape and the first move; the examples then walk each cell.
Every problem below is a cell in this grid. "Direct" means L'Hôpital applies with no setup; "Convert" means you must reshape it into a fraction first.
Cell
Form
First move (WHY)
Example
A
0 0 direct
already a fraction of two zeros
(1)
B
0 0 repeated
new limit is still 0 0 , apply again
(2)
C
∞ ∞ direct
growth race, compare speeds
(3)
D
0 ⋅ ∞ convert
rewrite f ⋅ g = 1/ g f
(4)
E
∞ − ∞ convert
common denominator kills the subtraction
(5)
F
1 ∞ (log)
take ln , drop the exponent
(6)
G
0 0 (log)
take ln , exponent becomes a product
(7)
H
trap: looping
rule cycles forever → use algebra
(8)
I
word problem
build the limit yourself, then solve
(9)
J
exam twist: one-sided sign
the sign of the approach decides ± ∞
(10)
Notation reminder, earned before use:
x → a lim reads "the value x 's expression heads toward as x creeps to a " — see Limits — Definition & Laws .
f ′ ( x ) is the derivative : the instantaneous rate of change, the slope of the tangent line — see Linear Approximation & Tangent Lines .
x → 0 + means x approaches 0 from the right (only positive values); x → 0 − from the left.
x → 0 lim x 3 tan x − x
Forecast: guess before reading — is it 0 , a finite nonzero number, or ∞ ? (The subtraction on top hints the numerator is tiny , tinier than x but maybe comparable to x 3 .)
Check the form. At x = 0 : tan 0 − 0 = 0 on top, 0 3 = 0 on bottom. So 0 0 . ✔
Why this step? You may never apply L'Hôpital without confirming both parts hit 0 (or both ∞ ).
Differentiate top and bottom separately. d x d ( tan x − x ) = sec 2 x − 1 , and d x d x 3 = 3 x 2 . New limit: 3 x 2 sec 2 x − 1 .
Why this step? The parent's proof says the ratio equals the ratio of speeds; that's g ′ f ′ , not the quotient rule.
Still 0 0 ? At 0 : sec 2 0 − 1 = 1 − 1 = 0 , bottom 0 . Yes → apply again. d x d ( sec 2 x − 1 ) = 2 sec 2 x tan x , bottom 6 x : 6 x 2 sec 2 x tan x .
Why? A fresh 0 0 is a fresh licence to reuse the rule.
Still 0 0 . Split: 6 x 2 sec 2 x tan x = 3 sec 2 x ⋅ x tan x . We know x t a n x → 1 and sec 2 0 = 1 , so limit = 3 1 .
Why factor instead of a third L'Hôpital? Recognising a known limit is cleaner and avoids messy derivatives.
Verify: Taylor gives tan x = x + 3 x 3 + ⋯ , so tan x − x ≈ 3 x 3 , and dividing by x 3 gives 3 1 . ✔ (see Taylor Series )
x → 0 lim x 2 e x − 1 − x
Forecast: the top is "e x minus its own linear approximation" — what's left after removing the line? Guess the value.
Form: e 0 − 1 − 0 = 0 ; bottom 0 . 0 0 ✔.
First L'Hôpital: top′ = e x − 1 , bottom′ = 2 x → 2 x e x − 1 . Why? Same speed-ratio principle.
Still 0 0 (e 0 − 1 = 0 ). Apply again: 2 e x → 2 e 0 = 2 1 .
Why repeat? Each new 0 0 resets the rule; stop only when you get a determined value.
Verify: e x = 1 + x + 2 x 2 + ⋯ , so e x − 1 − x ≈ 2 x 2 ; over x 2 that's 2 1 . ✔
Look at the figure: it plots the numerator and denominator both racing to ∞ , but at wildly different rates.
x → ∞ lim e x x 2
Forecast: which wins — a polynomial or the exponential? Guess 0 , 1 , or ∞ .
Form: as x → ∞ , top → ∞ , bottom → ∞ . ∞ ∞ ✔.
First L'Hôpital: e x 2 x . Still ∞ ∞ .
Why? Differentiating knocks the polynomial's power down by one each time, but the exponential never shrinks — see Exponential & Logarithm Growth Rates .
Second: e x 2 → 0 . Why stop? Numerator is now a constant, denominator → ∞ , giving a determined 0 .
Verify: the red curve e x in the figure overtakes x 2 and pulls away — the ratio flattens to the x -axis. Numerically at x = 10 : e 10 100 ≈ 0.0045 . ✔
x → 0 + lim x ln x
Forecast: x → 0 but ln x → − ∞ . A tug of war — who wins?
Form: 0 ⋅ ( − ∞ ) . Not a fraction, so L'Hôpital cannot touch it yet.
Convert. Put the "∞ " piece on top: x ln x = 1/ x ln x = x − 1/2 ln x , now ∞ − ∞ .
Why this split? Placing ln x on top gives an easy derivative 1/ x ; the other split (x on top over 1/ ln x ) makes a horrible denominator. Choose the split that simplifies.
L'Hôpital: top′ = x 1 , bottom′ = − 2 1 x − 3/2 . Ratio = − 2 1 x − 3/2 1/ x = − 2 x 1/2 = − 2 x → 0 .
Verify: at x = 0.0001 : 0.0001 ln ( 0.0001 ) = 0.01 × ( − 9.21 ) ≈ − 0.092 , already crawling toward 0 . ✔
x → 0 lim ( sin x 1 − x 1 )
Forecast: both pieces blow up to + ∞ near 0 . Their difference might be finite — guess it.
Form: as x → 0 + , s i n x 1 → ∞ and x 1 → ∞ , so ∞ − ∞ . Indeterminate.
Convert to one fraction (common denominator): x sin x x − sin x .
Why? A single fraction turns the subtraction into a clean 0 0 (top → 0 , bottom → 0 ), which the rule can eat.
L'Hôpital: top′ = 1 − cos x , bottom′ = sin x + x cos x → sin x + x cos x 1 − cos x . Still 0 0 .
Again: top′ = sin x , bottom′ = 2 cos x − x sin x → 2 cos x − x sin x sin x → 2 0 = 0 .
Verify: Taylor: x − sin x ≈ 6 x 3 and x sin x ≈ x 2 , ratio ≈ 6 x → 0 . ✔
x → 0 + lim ( cos x ) 1/ x 2
Forecast: base → 1 , exponent → ∞ . A 1 ∞ — is the answer 1 , e , or something less than 1 ?
Form: cos 0 = 1 , exponent 1/ x 2 → ∞ . 1 ∞ — indeterminate because "1 -ish raised to huge" could be anything.
Take ln . Let y = ( cos x ) 1/ x 2 , then ln y = x 2 ln ( cos x ) .
Why ln ? Exponents are invisible to differentiation until ln drags them down to a product/quotient — see parent Example (5).
Form of ln y : ln ( cos 0 ) = ln 1 = 0 , over x 2 = 0 : 0 0 . L'Hôpital: top′ = cos x − sin x = − tan x , bottom′ = 2 x → 2 x − tan x → − 2 1 (using tan x / x → 1 ).
Undo the log. ln y → − 2 1 ⇒ y → e − 1/2 = e 1 .
Why exponentiate? We solved for ln y ; the original limit is e ( l n y ) .
Verify: e − 1/2 ≈ 0.6065 . At x = 0.1 : ( cos 0.1 ) 100 = ( 0.995004 ) 100 ≈ 0.6062 . ✔
x → 0 + lim x x
Forecast: base → 0 (pulls toward 0 ), exponent → 0 (pulls toward 1 ). Who wins?
Form: 0 0 — indeterminate tug of war.
Take ln . y = x x ⇒ ln y = x ln x . That is Cell D, form 0 ⋅ ( − ∞ ) .
Convert & solve. x ln x = 1/ x ln x , form ∞ − ∞ . L'Hôpital: − 1/ x 2 1/ x = − x → 0 .
Why the 1/ x split? Same reason as Example 4 — keeps ln x 's easy derivative on top.
Undo the log. ln y → 0 ⇒ y → e 0 = 1 .
Verify: at x = 0.01 : 0.0 1 0.01 = e 0.01 l n 0.01 = e − 0.046 ≈ 0.955 , climbing toward 1 . ✔
x → ∞ lim x x 2 + 1
Forecast: for huge x , x 2 + 1 ≈ x . Guess the value.
Form: ∞ ∞ ✔ — so L'Hôpital is allowed , but watch what happens.
Try L'Hôpital: top′ = x 2 + 1 x , bottom′ = 1 → x 2 + 1 x . This is the reciprocal of the original — applying again just flips it back. Infinite loop.
Why does it loop? The rule requires the RHS limit to exist and simplify ; here it never simplifies.
Fix with algebra. Divide inside the root by x : x x 2 + 1 = x 2 x 2 + 1 = 1 + x 2 1 → 1 = 1 .
Why algebra beats the rule here? Factoring reveals the hidden structure the derivative kept hiding.
Verify: at x = 1000 : 1000001 /1000 ≈ 1.0000005 . ✔
Common mistake "The rule always works"
Wrong belief: if the form is ∞ ∞ , keep hammering L'Hôpital.
Why it feels right: the form check passed. Fix: if the ratio cycles or refuses to simplify, the RHS limit is not being reached — stop and use algebra .
Worked example Continuous compounding
A bank pays annual rate r but compounds n times a year, so $1 grows to ( 1 + n r ) n after one year. What happens as compounding becomes continuous , n → ∞ ?
Forecast: infinite compounding — does your dollar explode to ∞ , or settle at a finite number?
Set up the limit. We want n → ∞ lim ( 1 + n r ) n .
Why? "Continuous" = the number of compounding steps → ∞ ; that is the limit.
Form: base → 1 , exponent → ∞ : 1 ∞ (Cell F). Take ln : ln y = n ln ( 1 + n r ) , form ∞ ⋅ 0 .
Convert: ln y = 1/ n ln ( 1 + r / n ) , 0 0 . L'Hôpital in n : top′ = 1 + r / n − r / n 2 , bottom′ = − n 2 1 . Ratio = 1 + r / n r → r .
Undo log. ln y → r ⇒ y → e r .
Why exponentiate? The dollar value is y , not ln y .
Verify: with r = 1 this is the parent's Example (5), giving e 1 = e ≈ 2.71828 . Units: dollars per dollar (dimensionless growth factor). ✔
x → 0 ± lim x 3 sin x — treat both sides
Forecast: it's 0 0 , but the answer might not be a single number. Why might the two sides disagree?
Form: sin 0 = 0 , 0 3 = 0 : 0 0 ✔.
L'Hôpital: 3 x 2 cos x . Now top → cos 0 = 1 = 0 , bottom → 0 + (since x 2 > 0 for either sign of x ).
Why note the bottom's sign? 3 x 2 is always positive, so the fraction is + tiny + 1 = + ∞ from both sides.
Both-sided conclusion: x → 0 lim x 3 sin x = + ∞ .
Why the twist matters: had the denominator been x 3 without the L'Hôpital step, its sign would flip (x 3 < 0 left, > 0 right) — but after one differentiation the denominator became x 2 , which is sign-blind. Always recheck the sign after each step.
Contrast case. For x → 0 lim x 2 sin x : L'Hôpital gives 2 x cos x , and 2 x does flip sign: → + ∞ from the right, → − ∞ from the left → the two-sided limit does not exist .
Verify: at x = ± 0.01 : sin ( 0.01 ) / ( 0.01 ) 3 ≈ 0.01/1 0 − 6 = 1 0 4 > 0 both sides. ✔ And sin ( 0.01 ) / ( 0.01 ) 2 ≈ + 100 , while sin ( − 0.01 ) / ( − 0.01 ) 2 ≈ − 100 — opposite signs confirm no two-sided limit. ✔
Recall Cover the answers
Before L'Hôpital, what must you always do? ::: Confirm the form is 0 0 or ∞ ∞ .
lim x → ∞ x 2 / e x = ? ::: 0 (exponential beats any polynomial).
lim x → 0 + x x = ? ::: 1 .
lim x → 0 + ( cos x ) 1/ x 2 = ? ::: e − 1/2 .
When the rule loops (Example 8), what do you do? ::: Abandon it and use algebra.
Why can lim x → 0 sin x / x 2 fail to exist while sin x / x 3 gives + ∞ ? ::: After L'Hôpital the denominators are 2 x (sign flips) vs 3 x 2 (always positive).
Mnemonic The five first-moves
"Direct, Repeat, Convert, Log, or Algebra." Ask in that order: is it already 0 0 /∞ ∞ ? still after one step? need a fraction (0 ⋅ ∞ , ∞ − ∞ )? need ln (1 ∞ , 0 0 , ∞ 0 )? or is it looping (use algebra)?
L'Hôpital's rule — proof using linear approximation, 0 - 0, ∞ - ∞, other indeterminate forms (index 4.1.26) — parent: the rule and its proof.
Indeterminate Forms — the seven shapes of trouble enumerated in the matrix.
Taylor Series — every "Verify" via leading terms lives here.
Exponential & Logarithm Growth Rates — Examples 3, 4, 6, 7 are growth races.
Limits — Definition & Laws — one-sided limits and ± ∞ (Example 10).
Linear Approximation & Tangent Lines — why f ′ / g ′ is the right ratio.