Approximate area under a curve with left, right, midpoint, trapezoidal, and Simpson-style methods.
Results are calculated automatically as you enter data.
| Method | Estimate | Error vs reference | Note |
|---|
▼ See explanations and tips below ▼
A Riemann sum is a way to approximate a definite integral by breaking an interval into smaller pieces and adding up simple area estimates over those pieces. Instead of finding the exact area under a curve all at once, you divide the interval from \(a\) to \(b\) into subintervals, choose a sample point in each subinterval, evaluate the function there, and multiply by the subinterval width.
For a function \(f(x)\), a basic Riemann sum has the form:
Here, \(x_i^*\) is the chosen sample point in the \(i\)th subinterval, and \(\Delta x_i\) is the width of that subinterval.
Riemann sums are important because they connect the geometric idea of area under a curve with the formal definition of the definite integral. When the subintervals get smaller and the sum approaches a limiting value, that limit is the definite integral:
In practice, Riemann sums and related numerical integration rules are useful when an exact antiderivative is hard to find, unnecessary, or unavailable. They are also useful for checking work, visualizing accumulation, and comparing how different approximation methods behave.
Numerical integration gives you a practical way to estimate a definite integral using function values. That matters in calculus because not every useful integral has a simple exact form, and many real-world situations start with sampled data or a function that is easier to evaluate than to integrate symbolically.
Students use Riemann sums to understand what an integral means. Instructors use them to show how left endpoints, right endpoints, midpoints, trapezoids, and Simpson's rule can produce different estimates from the same function. Scientists, engineers, and analysts use the same general idea when they estimate accumulated change, area, distance, mass, probability, work, or other quantities represented by integrals.
For equal-width subintervals, the interval from \(a\) to \(b\) is divided into \(n\) pieces. The width of each piece is:
The endpoints are:
where \(i = 0, 1, 2, \ldots, n\).
The simplest Riemann sum methods use rectangles. Each rectangle has width \(h\), and its height is a function value chosen from the corresponding subinterval.
The left endpoint rule uses the left side of each subinterval:
The right endpoint rule uses the right side of each subinterval:
The midpoint rule uses the middle of each subinterval:
The trapezoidal rule averages the function values at the two endpoints of each subinterval:
Equivalently, it can be written as:
Simpson's rule uses a weighted pattern of endpoint and interior values. For an even number of panels, it is:
Simpson's rule needs an even number of panels because it works in pairs of subintervals. If an odd panel count is entered for Simpson's rule, this calculator adjusts it to the next even number before calculating the Simpson estimate.
Suppose you want to approximate:
Use \(n=4\) equal subintervals:
The subinterval endpoints are:
The left endpoint estimate is:
The right endpoint estimate is:
The midpoint estimate uses the midpoints \(0.25\), \(0.75\), \(1.25\), and \(1.75\):
The exact value of this integral is:
This example shows why the choice of sample point matters. With only four subintervals, the left and right endpoint estimates are far apart, while the midpoint estimate is closer.
For the same integral, the trapezoidal rule with \(n=4\) gives:
Simpson's rule with \(n=4\) gives:
In this case, Simpson's rule gives the exact value because the function is a simple polynomial that fits the rule especially well. In general, Simpson's rule is still an approximation, not a guarantee of exactness for every function.
Consider:
This integral is positive because \(\sin(x)\) is above the \(x\)-axis on most of the interval from \(0\) to \(\pi\). Its exact value is:
If the bounds are reversed, the sign changes:
This is not a calculator error. A definite integral is signed, and reversing the interval reverses the sign. The same idea explains why an integral can be negative when a function lies below the \(x\)-axis.
The selected estimate is the calculator's numerical approximation of the signed definite integral over the interval you entered. It is not automatically the exact value of the integral.
A positive estimate usually means the positive signed area dominates over the interval. A negative estimate can happen when the function is mostly below the \(x\)-axis or when the interval is entered in reverse order. A result near zero may mean the positive and negative signed areas nearly cancel, not necessarily that there is no area under the curve.
The delta x value is the panel width used by the selected method. A smaller absolute value of \(\Delta x\) means the interval has been divided into more panels. For smooth functions, using more panels often improves the approximation, although difficult functions can still require caution.
The reference value is computed on independently refined composite Gauss–Legendre sample grids: one dyadic and one non-nested. It is reported only when both grids converge and agree within tolerance. It is useful for comparison, but it is still numerical and should not be treated as a symbolic or exact integral.
The error shown by the calculator is:
A positive error means the selected method is above the reference estimate. A negative error means it is below the reference estimate. Because the reference is approximate, this error is also an approximate comparison.
The calculator rejects intervals where its bounded numerical checks encounter a non-finite value or cannot resolve a possible discontinuity or sharp feature. Split such an interval at the problematic point before estimating each piece.
The best method label identifies which of the displayed methods is closest to the reference estimate for the current input. It does not prove that the method is always best for that function or that the reference is exact.
One common mistake is assuming that a Riemann sum is the same as the exact integral. A finite Riemann sum is an approximation. The definite integral is the limiting value approached as the subintervals become smaller under the right conditions.
Another common mistake is confusing signed area with total geometric area. If a function dips below the \(x\)-axis, the definite integral subtracts that area. To find total geometric area, you must account for where the function is positive and negative.
Implicit multiplication is accepted: 2x and 2*x both mean two times x. Explicit * can make longer expressions easier to read. Use \(x\) as the variable, and use supported function names such as sin(x), cos(x), sqrt(x), log(x), ln(x), and exp(x).
For trigonometric functions, inputs are interpreted in radians. For example, use sin(pi/2) for a sine value of \(1\), not sin(90).
For Simpson's rule, do not be surprised if an odd subinterval count is adjusted. Simpson's rule needs an even number of panels, so the calculator uses the next even panel count for that method.
Finally, avoid rounding too early when checking work by hand. Small rounding differences can become visible when many function values are added together.
Use Riemann sums or numerical integration when you need to:
These methods are especially helpful for learning because they show the relationship between function values, interval width, signed area, and approximation error.
Numerical integration depends on the function, the interval, the method, and the number of panels. A smooth, slowly changing function is usually easier to approximate than a function with sharp turns, jumps, vertical asymptotes, or rapid oscillation.
This calculator does not perform symbolic integration. It evaluates the function numerically at selected points and combines those values according to the chosen method.
The reference value is also approximate. The calculator checks independently refined, differently sized sample grids before showing it, but no bounded numerical check can prove an exact integral. Functions with discontinuities, singularities, unresolved sharp features, or very rapid oscillations may be rejected so you can split the interval and check each piece.
The calculator rejects equal lower and upper bounds, non-integer subinterval counts, subinterval counts below \(2\), and subinterval counts above \(10000\). It also rejects functions that produce non-finite values at evaluated points.
Graphing is a visualization aid, not proof of accuracy. The curve is sampled, and the visual display may use fewer panels than the calculation when many subintervals are entered.
No unit conversion is performed. If your \(x\)-values have units and your function values have units, the integral's units are the product of those units. For example, integrating velocity in meters per second with respect to seconds gives meters. If your inputs are abstract math values, the result is simply an abstract signed integral value.
For coursework, check your instructor's required method and rounding rules. For high-stakes engineering, scientific, financial, safety, or official work, verify the result with an appropriate numerical method and consult a qualified professional when needed.
2x or explicit multiplication such as 2*x; explicit * can make longer expressions easier to read. Use supported function names such as sin(x), sqrt(x), ln(x), and exp(x).
A finite Riemann sum is usually an approximation. The exact definite integral is related to the limit of Riemann sums as the subinterval widths approach zero, assuming the function is integrable. With a finite number of panels, the result depends on the method and sample points used.
There is no single method that is always best for every function and interval. For many smooth functions, midpoint, trapezoidal, and Simpson estimates improve as the number of panels increases, and Simpson's rule often performs well because it uses a weighted pattern of function values. The comparison table helps you see which method is closest to the calculator's reference estimate for the current input.
Composite Simpson's rule works by grouping subintervals in pairs and fitting a quadratic-style approximation over each pair. Because of that pairing, the number of panels must be even. If you enter an odd number for Simpson's rule, this calculator adjusts it to the next even number.
A negative estimate means the signed integral is negative. This can happen when the function is mostly below the \(x\)-axis or when the bounds are entered in reverse order. It does not necessarily mean the visual area is negative; it means the signed accumulation is negative.
The error is a comparison against the calculator's independently refined numerical reference, not a guaranteed exact mathematical error. It is useful for seeing whether the selected method is above or below that reference and by roughly how much. For difficult functions, split the interval at a suspected discontinuity or sharp feature and verify each piece independently.
No. Trigonometric inputs are evaluated in radians. To use common angle values, enter expressions such as pi/2, pi, or 2*pi rather than degree values like 90 or 180.
Books and Textbooks