Limit Calculator
Enter a function of x and the value x approaches. The calculator samples the function from both the left and the right, builds a table of values, and reports the limit when both sides agree — or that the limit does not exist when they differ.
The definition, carried out literally
A limit asks what a function approaches near a point, deliberately ignoring what happens at the point itself. That distinction is the whole idea: sin(x)/x is undefined at zero, division by zero being what it is, yet the values either side close in on 1 without ambiguity.
This calculator answers that question numerically. It evaluates the function at points marching in towards your target from the left and from the right, shows the table it built, and reports a limit only when the two sides agree.
Textbook limits are found by algebra — factor and cancel, apply an identity, use a standard result. This tool does something more elementary: it looks. Sampling ever closer is the definition performed rather than reasoned about, and the table shows the approach rather than asserting its endpoint.
That makes it an independent check on algebraic work, since the methods share nothing. It also makes the one-sided values visible, turning does-not-exist from a verdict into an observation.
How to use this calculator
- Enter the function of x Standard notation with ^ for powers and the usual named functions. The function does not need to be defined at the target point — that is frequently the whole reason for asking.
- Enter the point x approaches Any finite number. It is the value the samples close in on from both directions, not a value the function is evaluated at.
- Read the table before the conclusion The rows run from furthest to nearest on the left, then nearest to furthest on the right. Watching the values settle is more informative than the single number at the end.
- Compare the two one-sided lines They are reported separately. Agreement gives the limit; disagreement is itself the answer, and the size of the gap says what kind of discontinuity it is.
How the limit is estimated
The function is sampled at five distances either side of the target — a tenth, a hundredth, down to a hundred-thousandth — giving ten rows that sweep in from the left and back out to the right. Those rows are for reading; the verdict is not based on them.
The decision uses two closer probes instead, one a millionth below the point and one a millionth above. Those are the reported left-hand and right-hand values, and the limit is their average when they agree.
Agreement is judged relatively, not absolutely. The probes must differ by less than a ten-thousandth of the larger, so a function taking values in the millions is not called discontinuous over a difference negligible at its own scale. Both must be finite: an infinite or undefined probe means no limit is reported.
Because the probes sit a millionth away rather than at the point, a removable discontinuity is invisible to the method — which is exactly what makes it work. The hole at the centre is stepped over, and the surrounding values determine the answer.
What each input means
- f(x) Function — form field “Function f(x)”
- The expression to examine. It is compiled once and evaluated at each probe, and it may be undefined at the target point without that causing a problem.
- a Target point — form field “x approaches”
- The value x approaches, which must be a finite number. The function is never evaluated exactly here.
- L⁻, L⁺ One-sided values
- The function a millionth below and a millionth above the target. Their agreement is the test, and their average is the reported limit.
Worked examples
Every number below is produced by the same calculation engine the tool above runs. Nothing here is typed by hand, so the walkthrough cannot drift from what you get when you enter the same values yourself.
The standard limit of sin(x)/x
Take sin(x)/x as x approaches 0. The function is undefined there, but the limit is the well-known value 1 — the result the derivative of sine ultimately rests on.
Inputs Function f(x) = sin(x) / x, x approaches = 0
- Limit lim x→0 of sin(x) / x
- x = -0.00001 f(x) = 1
- x = -0.0001 f(x) = 1
- x = -0.001 f(x) = 1
- x = -0.01 f(x) = 0.999983
- x = -0.1 f(x) = 0.998334
- x = 0.1 f(x) = 0.998334
- x = 0.01 f(x) = 0.999983
- x = 0.001 f(x) = 1
- x = 0.0001 f(x) = 1
- x = 0.00001 f(x) = 1
- Left side x→0⁻ f(x) → 1
- Right side x→0⁺ f(x) → 1
- Conclusion Both sides agree → limit ≈ 1
Result lim x→0 ≈ 1
The table climbs towards 1 from below on both sides, since sine is slightly smaller than its argument for small non-zero values. Watching it approach convinces most people more than the squeeze-theorem proof does on first meeting.
Nothing in the output flags that the function has no value at 0. The probes sit either side of the gap, so the hole is simply never visited.
Two sides that disagree
Take abs(x)/x as x approaches 0. The expression is a sign function in disguise: minus one to the left of zero, plus one to the right.
Inputs Function f(x) = abs(x) / x, x approaches = 0
- Limit lim x→0 of abs(x) / x
- x = -0.00001 f(x) = -1
- x = -0.0001 f(x) = -1
- x = -0.001 f(x) = -1
- x = -0.01 f(x) = -1
- x = -0.1 f(x) = -1
- x = 0.1 f(x) = 1
- x = 0.01 f(x) = 1
- x = 0.001 f(x) = 1
- x = 0.0001 f(x) = 1
- x = 0.00001 f(x) = 1
- Left side x→0⁻ f(x) → -1
- Right side x→0⁺ f(x) → 1
- Conclusion Left and right values disagree — the limit does not exist.
Result lim x→0 does not exist
The two one-sided lines report opposite values, and the conclusion is that no two-sided limit exists. That is a definite mathematical statement, not a failure to converge — each side has a perfectly good limit of its own.
The table makes the shape of the failure clear. The values do not oscillate or blow up; they simply sit at two different constants, which is the signature of a jump discontinuity.
Reading the result
An estimate, not a proof
The answer is marked as approximate because it is measured. For a function that is well behaved near the point it is reliable to several figures, but no amount of sampling establishes a limit with certainty — only an argument does.
Does not exist is a real answer
When the two sides disagree the limit genuinely does not exist, and that is often the point of the question. Look at the two one-sided values to see whether the cause is a jump, a sign change, or a blow-up in one direction.
Where sampling can mislead
A function oscillating faster and faster near the point, such as sin(1/x) at zero, may return two probe values that happen to be close without the limit existing at all. The method tests two points, and two points cannot detect that.
When you would use this
Checking an indeterminate form
After resolving a zero-over-zero form by factoring or by a standard result, sampling near the point confirms the value independently. The two routes share no machinery, so agreement is meaningful.
Seeing what a discontinuity looks like
Comparing the two one-sided values distinguishes a removable hole from a jump from an asymptote, which is the classification most introductory courses ask for.
Assumptions and limitations
What this calculator assumes
- The target point is a finite number, and the function is defined at least in a small neighbourhood either side of it.
- The verdict rests on two probes a millionth from the point, not on the displayed table.
- Agreement is measured relative to the size of the values, not as a fixed absolute tolerance.
- A non-finite value on either side means no limit is reported.
Where it stops being the right tool
- Limits as x tends to infinity cannot be requested: the target must be a finite number.
- Rapid oscillation near the point can defeat a two-probe test entirely.
- The result is numerical throughout: no symbolic manipulation, no L'Hôpital's rule, and no exact value such as a fraction or a multiple of π.
Common mistakes
Expecting the function to be defined at the point
Why it happens. It seems reasonable that a limit at a point requires a value there. It does not, and the most instructive limits are precisely those where the value is missing.
How to avoid it. Ask the question anyway. A hole in the function is no obstacle, since the probes step over the point rather than onto it.
Reading does-not-exist as a computation failure
Why it happens. Every other result is a number, so a sentence in its place reads like the tool giving up rather than answering.
How to avoid it. Read the two one-sided lines. Two different finite values is a jump discontinuity, and the correct answer is that no two-sided limit exists.
Quoting the last digit as exact
Why it happens. A limit that comes out as 1 or 2 looks exact, and the noise-removal step makes clean values look cleaner still.
How to avoid it. Treat the figure as an estimate. Where the exact value matters, confirm it algebraically and use this only as corroboration.
Key terms
Frequently asked questions
How is the limit estimated?
By sampling. The function is evaluated at points closer and closer to the target from both sides, and the verdict comes from two probes a millionth either side of it. When those agree to within a ten-thousandth of their own magnitude, their average is reported as the limit.
Can it detect when a limit does not exist?
Yes, when the failure is a disagreement between the two sides. If the left-hand and right-hand probes differ, or either is non-finite, the tool reports that no two-sided limit exists rather than averaging them.
Does it handle removable discontinuities?
Yes, and they are the natural case for it. For a hole such as (x² − 1)/(x − 1) at x = 1 the probes sit either side of the missing point and converge on the value the cancelled expression would give.
Why is the answer marked approximate?
Because it is measured rather than derived. A numerical estimate from finitely many samples is close to the true limit for a well-behaved function, but it is never a proof — and for a rapidly oscillating function it can be wrong.