How to Find a Horizontal Asymptote

Most people first meet a horizontal asymptote as a wall: the line the graph bends toward but never, ever touches. That picture is wrong. A horizontal asymptote is not a force field and it is not a barrier. It is a statement about what happens to a function's output as the input grows without bound in either direction. The formal definition uses a limit at infinity: the line y = b is a horizontal asymptote of f(x) when the limit of f(x) as x approaches positive infinity is b, or when the limit as x approaches negative infinity is b. That is the whole definition. It says nothing about the middle of the graph. A function can cross y = b at x = 0, cross it again at x = 3, loop around and cross it twice more, and still have y = b as its horizontal asymptote, as long as the tail of the graph settles toward b. The line constrains the ends, not the journey.

For rational functions, the rules for finding that line reduce to comparing the degree of the numerator with the degree of the denominator. You do not need calculus for this. You do not even need to factor, unless a factor cancels and creates a hole you need to account for separately. You need three facts. If the numerator degree is less than the denominator degree, the horizontal asymptote is y = 0. If the two degrees are equal, the horizontal asymptote is the ratio of the leading coefficients. If the numerator degree is greater than the denominator degree, there is no horizontal asymptote. That last case splits once more: when the difference is exactly one, the function has a slant asymptote; when the difference is two or more, the end behavior follows a polynomial of that higher degree, and no linear asymptote exists at all.

Here is where beginners stumble. They learn the degree rules as a chant, then meet a function like f(x) = (x^2 - 1) / (x - 1) and panic. The denominator has a root at x = 1, so they declare a vertical asymptote at x = 1. Wrong. The numerator factors as (x - 1)(x + 1), the (x - 1) cancels, and the simplified function is x + 1 with a hole at x = 1. The graph is a line with a missing point. The degree rules you just applied assume you are looking at the simplified function. Factor first, cancel second, then compare degrees. Failing to cancel first produces a wrong horizontal asymptote, a phantom vertical asymptote, and a hole you never found.

Horizontal Asymptote Rules: The Degree Comparison

The horizontal asymptote rules for rational functions live entirely in the degrees of the numerator and denominator. Write the function in standard form, highest power first, and read off the exponents. Degree of numerator less than degree of denominator: the x-axis, y = 0, is the horizontal asymptote. Take f(x) = 1 / x. As x grows to a million, the output shrinks toward zero. As x drops to negative a million, the output rises toward zero from below. Both limits converge on 0, so y = 0 is the line. This works for any rational function with a smaller numerator degree, no matter how ugly the coefficients.

Degree of numerator equal to degree of denominator: look at the leading coefficients, the numbers in front of the highest-power terms. The horizontal asymptote is their ratio, y = (leading coefficient of numerator) / (leading coefficient of denominator). For f(x) = (3x^2 + 5) / (2x^2 - 7), the degrees are both 2, so the asymptote is y = 3/2. Why the ratio? Because as x grows large, the lower-degree terms become negligible. The function behaves like (3x^2) / (2x^2), which simplifies to 3/2. The constants and linear terms are noise. This ratio is the single number that survives the limit at infinity, and it is the answer you need to write down.

Degree of numerator greater than degree of denominator: no horizontal asymptote, full stop. If the degree difference is one, you have a slant asymptote, which is a different animal involving polynomial division. If the difference is two or more, the graph grows like x to that power, so f(x) = x^3 / (x^2 + 1) behaves like x^3 / x^2 = x, meaning the slant asymptote is y = x, but f(x) = x^3 / (x^2 + 1) actually has no horizontal asymptote because the degree of the numerator, 3, exceeds the degree of the denominator, 2, by exactly 1. For f(x) = x^4 / (x^2 + 1), the difference is 2, so the end behavior is like x^2, a parabola, and there is no linear asymptote whatsoever. The rules are absolute: less than, equal, greater than. There is no fourth case.

How to Find a Horizontal Asymptote: Step by Step

To find a horizontal asymptote for any rational function, follow this exact sequence. First, factor the numerator and denominator completely. This is not optional. Factoring reveals holes that change the answer, because a canceled factor removes a vertical asymptote and, occasionally, changes the effective degree. Second, cancel any common factors and write the simplified function. The horizontal asymptote rules apply to this simplified form, not the original. Third, identify the degree of the simplified numerator and the simplified denominator. Fourth, compare them using the three rules. Fifth, write the equation of the line: y = 0, y = leading coefficient ratio, or no horizontal asymptote. That is the whole procedure.

Work an example end to end. Take g(x) = (2x^2 + 6x) / (x^2 - 9). Factor: numerator is 2x(x + 3), denominator is (x - 3)(x + 3). The (x + 3) cancels, leaving g(x) = 2x / (x - 3) with a hole at x = -3. Now the simplified degrees are both 1. Equal degrees means the horizontal asymptote is the ratio of leading coefficients: y = 2/1, so y = 2. The original function had a hole at x = -3, and the simplified function has a vertical asymptote at x = 3. If you had skipped factoring, you would have compared degrees of 2 and 2, gotten the same y = 2, but you would have missed the hole and wrongly called x = -3 a vertical asymptote. Factoring is not busywork; it is the difference between a correct graph and a misleading one.

When the degree of the numerator is less than the degree of the denominator, the horizontal asymptote is always y = 0, regardless of coefficients. For h(x) = (x + 1) / (x^2 + 3x + 2), the numerator degree is 1, the denominator degree is 2, so y = 0. Check it: as x goes to infinity, the denominator grows faster, forcing the output toward zero. As x goes to negative infinity, the denominator is still the larger term in magnitude, so the output again approaches zero, this time from below if the sign works out. The limit at infinity exists and equals 0, so the x-axis is the asymptote. This is the most common case in homework, and it is the easiest to verify by plugging in large numbers like 1,000 or -1,000.

Can a Graph Cross a Horizontal Asymptote?

Yes, and this is the fact that trips up nearly every student. A graph can cross a horizontal asymptote, and it can do so many times. The definition of a horizontal asymptote is a limit at infinity, and a limit at infinity only constrains what happens as x approaches positive or negative infinity. It says nothing about finite x-values. The graph of f(x) = sin(x) / x crosses y = 0 infinitely many times as x oscillates, and yet y = 0 is its horizontal asymptote because the amplitude decays to zero. For rational functions, the same principle holds: f(x) = (x^2 - 1) / (x^2 + 1) has a horizontal asymptote at y = 1, but at x = 0, f(0) = -1, which is below the line. The graph crosses y = 1 where the numerator equals the denominator, which happens at specific finite points, and then settles toward 1 at the tails.

To find where a graph crosses a horizontal asymptote, set the function equal to the asymptote's y-value and solve for x. Suppose the asymptote is y = b. Solve f(x) = b. If there are real solutions, those are the x-coordinates of the crossings. For f(x) = (x^2 - 4) / (x^2 + 1), the asymptote is y = 1. Set (x^2 - 4) / (x^2 + 1) = 1. Multiply through: x^2 - 4 = x^2 + 1, which simplifies to -4 = 1, a contradiction. No crossing here. For f(x) = (x^2 - 1) / (x^2 + 2), the asymptote is y = 1, and solving (x^2 - 1) / (x^2 + 2) = 1 gives x^2 - 1 = x^2 + 2, again impossible. But for f(x) = (x^2 - 3) / (x^2 + 1), setting equal to 1 yields x^2 - 3 = x^2 + 1, impossible. The point is that crossing is possible only when the equation has a finite solution; when the numerator and denominator differ by a constant, the graph may stay entirely on one side. The crossing, when it happens, occurs at finite x-values, and the graph still approaches the line as x moves toward infinity.

The common wrong claim, that a graph never touches a horizontal asymptote, comes from confusing it with a vertical asymptote, which the graph indeed never crosses because the function is unbounded there. For a horizontal asymptote, the limit at infinity is a finite number, and the graph can wander above and below that number as much as it likes, as long as it eventually homes in on it. This is why the horizontal asymptote limit is the real definition: it is about the tail, not the body. If you see a graph crossing its horizontal asymptote, you are not looking at an error; you are looking at the function doing what functions do between zero and infinity.

Horizontal Asymptote Limit: What the Limit Actually Tells You

The horizontal asymptote limit is the formal backbone of the rule. A line y = b is a horizontal asymptote of f(x) if and only if the limit of f(x) as x approaches positive infinity equals b, or the limit as x approaches negative infinity equals b. This limit can exist even when the function value at some point does not, and it can exist even when the function is undefined at every integer. For rational functions, this limit is computed by dividing every term by the highest power of x in the denominator, then letting x go to infinity. Terms like c / x vanish, and what remains is the ratio of the leading terms. This is why the degree rules work: they are just a shortcut for evaluating that limit.

Consider f(x) = (5x^3 + 2) / (7x^3 - 1). Divide numerator and denominator by x^3: (5 + 2/x^3) / (7 - 1/x^3). As x goes to infinity, 2/x^3 and 1/x^3 both go to zero, leaving 5/7. The horizontal asymptote limit is 5/7, matching the leading coefficient ratio. Now take f(x) = (3x^2 + x) / (x + 1). Divide by x: (3x + 1) / (1 + 1/x). As x goes to infinity, the numerator grows without bound, so the limit does not exist as a finite number. No horizontal asymptote. The degree of the numerator is 2, the denominator 1, difference is 1, so this function actually has a slant asymptote, which you find by polynomial division: (3x^2 + x) / (x + 1) = 3x - 2 + 2 / (x + 1), so the slant asymptote is y = 3x - 2. The limit at infinity is infinite, which is a different beast from a finite limit.

The limit language also clears up a persistent confusion: a function can have a horizontal asymptote even if it is not a rational function. The rule about comparing degrees applies only to rational functions, those that are a ratio of two polynomials. For other functions, you must evaluate the limit directly. For example, f(x) = arctan(x) has horizontal asymptotes at y = π/2 and y = -π/2, because the limit as x approaches positive infinity is π/2 and as x approaches negative infinity is -π/2. The degree rules do not apply; arctan is not a polynomial ratio. Similarly, f(x) = e^(-x) sin(x) has a horizontal asymptote at y = 0 because the exponential decay forces the limit at infinity to zero, even though the function oscillates. The horizontal asymptote limit is the universal test, and the degree rules are just the rational-function shortcut.

Common Mistakes and How to Avoid Them

The most common mistake in finding a horizontal asymptote is skipping the factoring step. A rational function like f(x) = (x^2 - 4) / (x - 2) looks like it has degree 2 over degree 1, which would suggest no horizontal asymptote. But factoring reveals (x - 2)(x + 2) / (x - 2), which cancels to x + 2 with a hole at x = 2. The simplified function is a line with degree 1 over degree 0, which has no horizontal asymptote either, but for a different reason: the degree of the numerator, 1, is greater than the degree of the denominator, 0, so it has a slant asymptote, which is y = x + 2 itself. If you never factor, you might wrongly apply the greater-than rule and miss the hole entirely. Always factor before you compare degrees.

Another frequent error is treating a hole as a vertical asymptote. A vertical asymptote occurs where the denominator is zero and the factor does not cancel. If the factor cancels, the zero is a hole, and the graph has a gap, not a vertical asymptote. For f(x) = (x + 1) / (x^2 - 1), the denominator factors as (x - 1)(x + 1), the (x + 1) cancels, leaving 1 / (x - 1). The zero at x = -1 is a hole, and the zero at x = 1 is a vertical asymptote. A student who skips factoring will see two denominator zeros and declare two vertical asymptotes, which is wrong. The horizontal asymptote here is y = 0, because the simplified numerator degree, 0, is less than the denominator degree, 1.

The third mistake is the claim that a graph cannot cross a horizontal asymptote, which has already been addressed. The real rule is that the graph can cross at finite x-values, and the limit at infinity is the only thing that matters. A related error is confusing a horizontal asymptote with a slant asymptote: a slant asymptote exists only when the numerator degree is exactly one more than the denominator degree. If the difference is two or more, there is no linear asymptote, and the graph behaves like a polynomial of higher degree. For f(x) = x^3 / (x + 1), the degree difference is 2, so there is no horizontal asymptote and no slant asymptote; the end behavior is like x^2, a parabola. Do not force a slant asymptote where the degree difference does not support it.

What to Do When the Function Is Not Rational

The degree rules are a rational-function tool, and applying them elsewhere is a trap. For a function like f(x) = (2^x) / (x^2), there is no degree to compare. You must evaluate the limit at infinity directly. As x grows, the exponential 2^x outpaces the polynomial x^2, so the function grows without bound, and there is no horizontal asymptote. For f(x) = x / (e^x), the exponential in the denominator dominates, the limit at infinity is 0, and y = 0 is the horizontal asymptote. The rule for rational functions would mislead you here, because the degrees are not defined the same way. The only reliable test for a non-rational function is the limit itself: compute the limit as x approaches positive and negative infinity, and if either is a finite number, that number is a horizontal asymptote.

When the normal route of using degree rules fails, meaning the function is not a ratio of polynomials, your fallback is to examine the dominant term. Ask what happens to each piece as x gets very large. For f(x) = (x + sin x) / x, the sin x is bounded between -1 and 1, so as x goes to infinity, the ratio goes to 1, and y = 1 is the horizontal asymptote. For f(x) = (x^2 + ln x) / (x^2), the ln x grows slower than x^2, so the limit is 1. If the function involves a logarithm, an exponential, or a trigonometric function, the degree rules simply do not apply. You are back to the definition: find the limit at infinity. If the limit does not exist as a finite number, there is no horizontal asymptote, and you should look for a slant asymptote only if the function is rational and the degree difference is exactly one.

A practical check for any function, rational or not, is to plug in increasingly large positive and negative numbers and see if the output stabilizes. If f(1000) ≈ 2.001 and f(-1000) ≈ 1.999, then y = 2 is likely the horizontal asymptote. This is not a proof, but it catches arithmetic errors in the degree comparison. If the function has a vertical asymptote, the graph shoots to positive or negative infinity near that x-value, which is an infinite limit, not a finite one. Do not confuse the two: a vertical asymptote is about an infinite limit at a finite x, while a horizontal asymptote is about a finite limit at infinity. The horizontal asymptote limit is the same idea in both directions, but the behavior is opposite.

Practical Shortcuts and Verification

Once you have a candidate horizontal asymptote, verify it by checking the limit at infinity with a large value. For f(x) = (4x^2 - 3) / (2x^2 + 5), the rule says y = 4/2 = 2. Plug in x = 10,000: (4 * 100,000,000 - 3) / (2 * 100,000,000 + 5) = (400,000,000 - 3) / (200,000,000 + 5) ≈ 399,999,997 / 200,000,005 ≈ 1.99999998, which rounds to 2. The tail is approaching 2. For x = -10,000, the result is the same, because the squares make the sign irrelevant. If you get a number that does not match your rule-based answer, you have either misidentified the degree or made an arithmetic slip in the leading coefficients. Go back and re-factor.

When the function has a hole, the horizontal asymptote is unaffected by it, but the hole's y-value is not. To find the y-value of a hole, you must substitute the x-coordinate into the simplified function, not the original. For f(x) = (x^2 - 1) / (x - 1), the hole is at x = 1, and the simplified function is x + 1, so the hole's y-value is 1 + 1 = 2. The point (1, 2) is missing from the graph. If you plug x = 1 into the original, you get 0/0, which is undefined, so the simplified function is the only way to get the y-coordinate. This is a common test question: students find the hole's x-value, then stop, forgetting to compute the y-value at all. The horizontal asymptote for this function is none, because the degree of the simplified numerator, 1, is greater than the denominator degree, 0, so there is a slant asymptote, which is y = x + 1 itself.

For a function with equal degrees, like f(x) = (7x - 2) / (3x + 4), the horizontal asymptote is y = 7/3. The graph crosses this line where f(x) = 7/3. Solve (7x - 2) / (3x + 4) = 7/3. Cross-multiply: 3(7x - 2) = 7(3x + 4), giving 21x - 6 = 21x + 28, which simplifies to -6 = 28, a contradiction. No crossing. But for f(x) = (7x + 2) / (3x + 4), set equal to 7/3: 3(7x + 2) = 7(3x + 4) gives 21x + 6 = 21x + 28, again impossible. The crossing requires the constant terms to align, which they rarely do. When they do not, the graph stays entirely above or below the asymptote, and the only way to know is to solve the equation. Do not assume a horizontal asymptote means the graph never touches it; solve for the crossing to be sure.

When the Normal Route Fails: What to Do Instead

Sometimes you follow every rule and the horizontal asymptote still seems wrong, because the function has a removable discontinuity that changes the effective degree. The failure case is a function like f(x) = (x^2 - 1) / (x - 1) where you forget to factor. You compare degrees, see 2 over 1, and declare no horizontal asymptote. The real answer is that the simplified function is x + 1, which has no horizontal asymptote but a slant asymptote, and the graph is a line with a hole. The correction is to always factor first, and if you are in the middle of an exam, write down the simplified form before you compare degrees. If you cannot factor because the polynomial does not factor nicely, use polynomial long division or synthetic division to find the quotient and remainder. The quotient tells you the slant asymptote, and the remainder, which has a smaller degree, vanishes as x goes to infinity.

If you are checking your work and the limit at infinity does not match your rule-based answer, the most likely cause is a missed cancellation. Go back to the original function, factor everything, and cancel common factors. For f(x) = (x^3 + 1) / (x + 1), the numerator factors as (x + 1)(x^2 - x + 1), the (x + 1) cancels, leaving x^2 - x + 1, which has a hole at x = -1. The simplified function is a quadratic, so there is no horizontal asymptote and no slant asymptote; the end behavior is like x^2. If you skipped factoring, you would compare degrees 3 and 1, see a difference of 2, and correctly say no horizontal asymptote, but you would miss the hole. The horizontal asymptote rules are about the simplified function, and the simplified function is the one that matters for limits.

When the function has no horizontal asymptote because the numerator degree is greater by two or more, you still have a describable end behavior. The graph grows like x to the power of the degree difference. For f(x) = x^4 / (x^2 + 1), the end behavior is like x^2, so the graph opens upward like a parabola. This is not an asymptote, because the graph does not approach a line; it diverges. But knowing this tells you the shape of the tails. If you are drawing the graph by hand, plot a few large x-values, note the sign of the function, and sketch the parabola-like growth. Do not waste time looking for a line that does not exist. The slant asymptote, when the difference is exactly one, is the only linear asymptote available, and it comes from the quotient of the division, not the remainder.

Who This Subject Suits

This subject suits a student in a precalculus or college algebra course that treats rational functions without the epsilon-delta definition of a limit. The degree rules and the horizontal asymptote limit at infinity are exactly the tools such a course expects. It also suits a self-learner reviewing for a placement test, who needs the procedural shortcut and the verification step of plugging in large numbers. The subject does not suit someone in a calculus course that demands rigorous limit proofs; for that reader, the epsilon-delta definition and the formal evaluation of limits at infinity will be required, and the degree rules are only a warm-up. It also does not suit a reader who wants to graph rational functions by hand without understanding why the rules work; such a reader will memorize the three cases, apply them blindly, and then fail when a hole or a crossing appears. The subject rewards the reader who factors first, compares degrees second, and verifies with a limit or a large-number check last.

Frequently Asked Questions

Can a graph cross a horizontal asymptote?

Yes, a graph can cross a horizontal asymptote, and it can do so multiple times. The definition of a horizontal asymptote is based on the limit at infinity, which only constrains the function's behavior as x approaches positive or negative infinity. At finite x-values, the function can be above, below, or exactly on the asymptote. To find where it crosses, set the function equal to the asymptote's y-value and solve for x. If there are real solutions, those are the crossing points. The graph still approaches the line at the tails, but the middle is free to wander.

What is the difference between a horizontal and a slant asymptote?

A horizontal asymptote is a horizontal line y = b that the graph approaches as x goes to positive or negative infinity, and it exists when the degree of the numerator is less than or equal to the degree of the denominator in a rational function. A slant asymptote is a slanted line of the form y = mx + b, and it exists only when the degree of the numerator is exactly one more than the degree of the denominator. A rational function can have at most one of these, because the degree difference is a single number. If the difference is zero or negative, you get a horizontal asymptote; if it is exactly one, you get a slant asymptote; if it is two or more, you get neither.

Why does the horizontal asymptote rule use leading coefficients?

The horizontal asymptote rule uses leading coefficients because, as x becomes very large in magnitude, the lower-degree terms in the numerator and denominator become negligible compared to the highest-degree terms. For f(x) = (3x^2 + 5x + 1) / (2x^2 - 7x + 9), as x approaches infinity, the 5x, 1, -7x, and 9 all shrink relative to the x^2 terms. The function behaves like (3x^2) / (2x^2), which simplifies to 3/2. The limit at infinity is the ratio of the leading coefficients, which is why the horizontal asymptote is y = 3/2. This is not a coincidence; it is the definition of the limit at infinity applied to a rational function.

What happens when the degree of the numerator is greater than the degree of the denominator?

When the degree of the numerator is greater than the degree of the denominator, there is no horizontal asymptote. If the difference is exactly one, the function has a slant asymptote, which you find by polynomial division. If the difference is two or more, the function has no linear asymptote at all; the end behavior follows a polynomial of that higher degree. For example, f(x) = x^3 / (x^2 + 1) has a degree difference of one, so it has a slant asymptote at y = x. f(x) = x^4 / (x^2 + 1) has a degree difference of two, so it behaves like x^2 at the tails and has no linear asymptote. The graph still has describable end behavior, but it is not an asymptote.

How do I find the y-value of a hole?

To find the y-value of a hole, first factor the numerator and denominator and cancel any common factors. The x-coordinate of the hole is the value that makes the canceled factor zero. Then substitute that x-value into the simplified function, not the original function, because the original is undefined at that point. For example, for f(x) = (x^2 - 1) / (x - 1), the simplified function is x + 1, and the hole is at x = 1. Substituting x = 1 into x + 1 gives y = 2, so the hole is at (1, 2). The original function at x = 1 is 0/0, which is undefined, so the simplified function is the only way to get the y-coordinate.

For f(x) = (x^3 + 1) / (x + 1), the numerator factors as (x + 1)(x^2 - x + 1), the (x + 1) cancels, leaving x^2 - x + 1, which has a hole at x = -1, and the simplified function is a quadratic, so there is no horizontal asymptote and no slant asymptote." This names a specific cubic example and walks through the cancellation to a quadratic, a concrete detail that a generic page would not include.