eduKate Singapore

VIEW THIS AS

Auto mode follows the Route Engine until you choose a viewpoint.

YOU ARE HERE

ROUTE CHECK

CONNECTED TO

WHAT NEXT

Use the canonical route for this room, or HELP if you are unsure.

Next useful routeCanonical hubHELP

Mathematics in Bukit Timah | The Trapezoidal Rule, Curved Areas and Numerical Estimates

—

by

eduKate Asia
Original editorial engineering illustration of a steady electricity load across a day versus a concentrated high peak with equal areas under both profiles.

A fictional garden has a smoothly curved upper boundary. A student knows its height at five evenly spaced positions, but not the exact shape between those measurements. Multiplying one length by one height would treat the region as a rectangle, while multiplying half the width by the maximum height would impose an unproven triangle. How can a set of measured heights be turned into a defensible area estimate? One answer is the trapezoidal rule.

This Mathematics in Bukit Timah guide uses imaginary cross-sectional profiles to explain trapeziums, numerical integration, unequal intervals, estimated areas, approximation error and the difference between a curve and its sampled data. Every garden, measurement, roofline, depth and rate in the examples is invented for education, not a survey of Bukit Timah Nature Reserve, Singapore Botanic Gardens or a real private property.

The immediate example has horizontal positions 0, 2, 4, 6 and 8 metres, with fictional vertical heights 0, 3, 4, 3 and 0 metres. Joining each adjacent pair of sampled points by straight lines creates four trapeziums. Their total area is 20 m². That value is an estimate of a smooth unknown curve, not automatically its exact enclosed area.

Article contents

Original graph illustration showing different curves with equal accumulated areas, used as context for numerical area estimation; no real Bukit Timah garden profile appears in the illustration.

Quick answer: average neighbouring heights, then multiply by width

A trapezium with two parallel vertical sides of lengths a and b separated by horizontal width h has area h(a+b)/2.

If an area under a graph is divided into intervals, estimate each curved section by a trapezium whose two vertical sides use the graph values at the interval’s endpoints.

Add all the trapezium areas to obtain the trapezoidal-rule estimate. The method replaces the unknown curved boundary between sample points with straight-line segments.

Why the Bukit Timah setting is fictional

Green areas and gardens around Bukit Timah provide an intuitive reason to ask about irregular boundaries. A classroom can imagine a curved plot without disturbing a physical location.

But no measured area, elevation, garden shape or access advice is supplied here. The profile numbers are defined on paper, and the resulting calculations apply only to that specified fictional model.

For actual land surveying or design, an appropriate qualified process would need real data, boundaries and uncertainty controls. The mathematical lesson is how one type of approximation works, not how to certify a real property area.

A straight-sided trapezium comes before a curved area

Imagine a shape with two parallel vertical sides measuring three and five metres, separated horizontally by four metres.

Its area is width × average parallel side: 4×(3+5)/2=16 m². This can be understood by rearranging or duplicating the trapezium into a rectangle or parallelogram with an equivalent area.

The trapezoidal rule uses that elementary shape repeatedly. There is no need to begin with calculus vocabulary before the learner understands the ordinary trapezium formula.

Worked example 1: one trapezium

A fictional graph has height y=2 m at x=0 and height y=6 m at x=3 m. Replace the boundary between those sample points with a straight line.

The horizontal width is three metres, and the vertical parallel sides are two and six metres. The trapezium area is 3×(2+6)/2=12 m².

If the true boundary is genuinely straight, the area is exact. If the actual curve bends above or below that connecting line, twelve is an approximation and may be too small or too large.

Curve samples do not determine a unique exact area

Five measured heights are useful but do not specify what the boundary did at every position between samples. Many different smooth curves can pass through exactly the same five points.

Some curves bow upward between points, while others dip. Their enclosed areas can differ even though the sample table is unchanged.

The trapezoidal rule makes an explicit interpolation assumption: between neighbouring sample points, use the straight connecting segment. The answer is exact for that piecewise-linear approximation, not for every possible unknown curve.

The five-point fictional garden profile

Use positions x=0,2,4,6,8 metres. Corresponding invented heights are f(x)=0,3,4,3,0 metres.

The four intervals are [0,2], [2,4], [4,6] and [6,8]. Each has equal width two metres.

All heights are nonnegative, so the area between the graph and horizontal axis can be estimated by adding ordinary nonnegative trapezium areas. State these conditions before beginning.

Worked example 2: calculate each trapezium separately

On [0,2], the two heights are 0 and 3. Area is 2(0+3)/2=3 m².

On [2,4], heights are 3 and 4. Area is 2(3+4)/2=7 m².

The remaining halves are symmetric: [4,6] contributes 7 m² and [6,8] contributes 3 m². Total is 3+7+7+3=20 m².

Why internal heights are counted twice

When adjacent trapeziums are added, each internal sample height contributes to the trapezium on its left and the trapezium on its right.

The two endpoint heights contribute only once, because each belongs to one outermost trapezium.

That counting pattern explains the familiar formula containing endpoint values once and internal values twice. Students should understand the reason instead of memorising a sequence of coefficients without a diagram.

The composite trapezoidal formula for equal widths

Suppose sample positions have common interval width h, and heights are y₀,y₁,…,yₙ at n+1 positions.

Then the composite trapezoidal estimate is T=(h/2)[y₀+2(y₁+⋯+yₙ₋₁)+yₙ].

The formula assumes equal spacing. If intervals have different widths, use the individual trapezium calculation for each interval instead of inserting an invented common h.

Worked example 3: use the composite formula

Our five heights are 0,3,4,3,0 and common h=2.

Calculate T=(2/2)[0+2(3+4+3)+0]. The internal sum is ten, so the bracket is twenty.

Therefore T=20 m², agreeing with the four separate trapeziums. Two independent layouts verify the arithmetic under the same approximation.

The subscript n counts intervals, not data points

Five sampled positions define four intervals. If a student uses n=5 as the number of trapeziums, the coefficient pattern may be applied incorrectly.

The subscript y₀ through y₄ contains five heights, with four steps between them. Each adjacent pair makes one trapezium.

This is a common off-by-one issue. A quick sketch marking four gaps beneath the five data points prevents it.

Worked example 4: count intervals before computing

A profile is observed at x=0,2,4,6,8. The points are five but the intervals are four.

The four trapezium contributions were 3,7,7,3. There is no fifth trapezium beyond x=8 because the domain ends there.

When more sample points are added inside the same domain, the number of intervals rises by one for each new interior point, while the total horizontal width stays unchanged.

The physical units depend on both axes

In the garden example, horizontal position is metres and vertical height is metres. Multiplying them produces square metres.

If the horizontal axis instead represented minutes and the vertical axis represented litres per minute, the area beneath a rate curve would have units of litres, not square metres.

Always multiply the units of the axes before interpreting a numerical “area under the graph”. The geometric trapezium formula stays the same, but the real-world meaning changes.

Worked example 5: rate multiplied by time

An invented classroom pump has a modelled flow rate of two litres per minute at t=0 and four litres per minute at t=5, with a straight-line rate change between those times.

The area beneath the rate–time graph is 5(2+4)/2=15 litres. That is the total volume transferred under the continuous-rate model.

No actual Bukit Timah water system is described. The exercise illustrates how an integral-like area can represent an accumulated amount with units derived from the axes.

The trapezoidal rule is exact for a straight line

If f(x) varies linearly over an interval, the straight segment connecting its two endpoints is the actual graph.

A trapezium constructed under that segment is therefore the true geometric region. No approximation error remains from replacing curvature with a straight line.

The rule becomes approximate when the graph bends between samples. Whether it overestimates or underestimates depends on the function’s curvature.

Worked example 6: a triangular graph

Take the line y=x/2 from x=0 to x=4. Its heights at the endpoints are zero and two.

The trapezoidal area is 4(0+2)/2=4 square units. The same region is a triangle with base four and height two, whose area is also 4.

This agreement is exact because the boundary is straight. Adding more sample points would produce the same total area.

A smooth quadratic curve lets us measure the approximation error

For a mathematical reference, define a fictional curved boundary f(x)=x(8−x)/4, valid on 0≤x≤8.

Its heights at x=0,2,4,6,8 are exactly 0,3,4,3,0, so it matches the original sampled table.

Unlike an unknown garden curve, we now know its complete equation. Advanced integration can calculate its exact area and compare it with the trapezoidal estimate of twenty.

Worked example 7: the exact quadratic area

Rewrite f(x)=2x−x²/4. An antiderivative is x²−x³/12.

Evaluate from zero to eight: at eight, 64−512/12=64−42⅔=64/3. At zero, the antiderivative is zero.

Thus exact signed area under this nonnegative quadratic on [0,8] is 64/3≈21.333 m². The four-trapezium estimate of 20 m² underestimates by 4/3≈1.333 m².

Why this particular trapezoidal estimate is too small

The reference parabola f(x)=x(8−x)/4 is concave downward. Between sampled positions, the curved graph lies above the straight chord joining those points.

Each trapezium therefore misses a little curved region above its straight top edge. Adding the missing pieces explains why the estimate is less than the exact area.

This reasoning applies to this function and interval. It is not a universal claim that trapezoidal estimates are always too small.

A convex curve can produce an overestimate

Consider f(x)=x² from x=0 to x=2. The endpoint heights are zero and four.

One trapezium over the interval has area 2(0+4)/2=4 square units. But exact integral area is 8/3≈2.667 square units.

The straight chord lies above the convex quadratic curve between the endpoints. In this case the trapezoidal rule overestimates by 4/3 square units.

Worked example 8: identify the sign of the error

For f(x)=x² on [0,2], the trapezoidal estimate is four and the exact integral is 8/3.

If error is defined as estimate minus exact, the signed error is 4−8/3=4/3, which is positive.

For our concave-down garden parabola, estimate minus exact is 20−64/3=−4/3, negative. An error calculation must state its sign convention before interpreting positive and negative values.

Smaller intervals can improve the estimate for a smooth curve

Keep the same quadratic f(x)=x(8−x)/4, but sample at every one metre instead of every two metres.

The heights for x=0 through 8 are 0,1.75,3,3.75,4,3.75,3,1.75,0.

The smaller intervals follow the curve more closely than the original longer chords. Recalculate the composite trapezoidal estimate to see how the error changes.

Worked example 9: eight trapeziums

With h=1, the seven internal heights sum to 1.75+3+3.75+4+3.75+3+1.75=21.

The trapezoidal estimate is (1/2)[0+2×21+0]=21 m².

The exact area remains 64/3≈21.333, so the underestimate is now just 1/3≈0.333 m². Halving the step width reduced the error by a factor of four in this quadratic example.

One coarse interval can be much worse

If we sample only at x=0 and x=8, both endpoint heights are zero. The single trapezium estimate would be 8(0+0)/2=0.

But the quadratic arch rises to four between the endpoints and encloses exact area about 21.333.

This extreme example shows why sampling frequency matters. A single pair of endpoint measurements can miss an entire hump that occurs between them.

Worked example 10: use two broad intervals

Instead, sample at x=0,4,8 with heights 0,4,0. Common h=4.

The two-trapezium estimate is (4/2)[0+2×4+0]=16 m².

Compared with the exact 64/3, the underestimate is 64/3−16=16/3≈5.333 m². This lies between the zero estimate from one interval and the more accurate estimates from four and eight intervals.

More samples do not guarantee improvement without conditions

For our smooth quadratic, smaller equal intervals improve accuracy predictably. But real data can contain measurement noise, abrupt changes or incorrect samples.

If an added observation is badly measured, including it might make a numerical estimate worse. Different sampling positions can also change error direction for complicated curves.

The safe principle is that finer, accurate sampling can improve approximation under suitable smoothness and measurement assumptions—not that adding arbitrary numbers guarantees truth.

Unequal intervals require individual widths

Suppose fictional heights are recorded at x=0,1,3,6, with values 0,2,4,1. The horizontal gaps are 1,2 and 3, not equal.

Do not use the equal-spacing composite formula with h=2 or h=3. Instead compute each trapezium separately.

The general expression is T=Σ[(xᵢ₊₁−xᵢ)(yᵢ+yᵢ₊₁)/2], summing over every neighbouring pair.

Worked example 11: calculate the unequal-width case

On [0,1], area is 1(0+2)/2=1. On [1,3], area is 2(2+4)/2=6.

On [3,6], area is 3(4+1)/2=7.5. Total is 1+6+7.5=14.5 square units.

A learner who divides the entire horizontal range by the wrong number of intervals and assumes equal spacing will generally obtain a different estimate from the supplied data.

Negative graph values change the interpretation

The trapezoidal rule applied algebraically to a function below the horizontal axis contributes negative signed area.

If the question asks for total geometric area between the curve and axis, below-axis regions must be counted positively, typically by splitting at zeros or working with absolute values.

For a physical garden height model, negative heights may mean the chosen baseline or model needs attention. A numerical formula cannot decide which interpretation is intended.

Worked example 12: signed area cancels

Consider the straight line f(x)=x−1 on [0,2]. At x=0 it is −1 and at x=2 it is +1.

One trapezoid-style signed calculation gives 2(−1+1)/2=0. The negative triangle from zero to one cancels the positive triangle from one to two.

But the geometric area between the line and horizontal axis is 1/2+1/2=1 square unit. The result depends on whether the question asks for signed accumulation or total region area.

The baseline must be defined

An area beneath a graph is measured relative to a chosen horizontal reference. In the garden example the reference is y=0 metres.

If heights instead describe elevation above an arbitrary datum, the area below the plotted values may not represent a meaningful physical ground surface without further context.

A strong word-problem solution names both axes, the baseline and any nonnegative assumptions before evaluating trapeziums.

Why an unknown curve still needs cautious reporting

Our original five height samples produced an estimate of twenty. If the true curve equation is not supplied, we cannot calculate the exact missed area from those samples alone.

We can compare numerical estimates under different assumed interpolations, or collect more measurements, but no calculator can reconstruct every possible curve that passes through the same points.

The word estimate should remain visible in the final answer. Writing 20.0000 m² does not make the underlying sampled curve known more accurately.

An illustrative upper bound from a stated constraint

Suppose a fictional profile is explicitly known to satisfy 0≤f(x)≤4 for every x from zero to eight. Its total geometric area must then lie between zero and 8×4=32 m².

The trapezoidal estimate twenty lies within that broad bound. The bound does not prove the estimate is close to the true area.

This illustrates the difference between a valid physical or mathematical constraint and an approximation procedure. Both can be useful, but they answer different questions.

Worked example 13: a conservative rectangle bound

Under 0≤f(x)≤4 across an eight-metre interval, enclosing rectangle area is 32 m².

No possible profile respecting the stated nonnegative height cap can enclose more than 32 m² over that domain.

The exact quadratic 64/3 and the trapezoidal estimate 20 both fit the bound. Yet infinitely many other allowable profiles could have very different areas.

The midpoint rectangle rule is a useful comparison

Another numerical method estimates each interval’s area from the height at its midpoint multiplied by the interval width.

For the reference quadratic with four intervals of width two, midpoints occur at x=1,3,5,7. Their heights are 1.75,3.75,3.75,1.75.

The midpoint rule produces a different estimate from endpoint-based trapeziums. Comparing the two methods can illuminate curvature and error direction.

Worked example 14: four midpoint rectangles

Add the four midpoint heights: 1.75+3.75+3.75+1.75=11.

Multiply by width two to obtain estimated area 22 m². The exact quadratic area is 64/3≈21.333 m².

The midpoint estimate overestimates by 22−64/3=2/3≈0.667 m², while the four-trapezium estimate underestimates by 4/3. The different methods bracket the exact result in this carefully chosen example.

Simpson’s rule is an advanced extension

When samples are equally spaced and the number of intervals is even, composite Simpson’s rule combines heights using alternating coefficients four and two.

For four equal intervals of width h, the rule is S=(h/3)[y₀+4y₁+2y₂+4y₃+y₄].

It fits quadratic behaviour differently from straight-line trapeziums. The formula’s requirements must be checked; it is not simply interchangeable with the trapezoidal rule on irregularly spaced data.

Worked example 15: Simpson’s rule on the reference parabola

With h=2 and heights 0,3,4,3,0, Simpson’s formula gives S=(2/3)[0+4(3)+2(4)+4(3)+0].

The bracket equals 12+8+12=32. Therefore S=64/3≈21.333 m².

This equals the exact area of the reference quadratic. Simpson’s rule is exact for polynomials up to degree three under its standard setup, including our degree-two example.

Do not apply Simpson’s rule with arbitrary spacing

If the sample positions are 0,1,3,6, the intervals are not equal. The straightforward composite formula just shown does not apply.

A student must either use a method designed for nonuniform intervals, resample with justified interpolation, or return to individual trapeziums.

A more complicated-looking formula is not automatically better if its mathematical assumptions are violated.

Estimation error can be quantified for sufficiently smooth curves

For a twice continuously differentiable function f on [a,b] with evenly spaced subinterval width h, a familiar composite trapezoid error bound is proportional to the maximum magnitude of f”.

One standard bound is |exact integral−T|≤(b−a)h²M/12 when |f”(x)|≤M throughout the interval.

This is an advanced guarantee with specific conditions. It should not be presented as a valid bound for an unknown jagged boundary or measurements with unmodelled errors.

Worked example 16: bound the quadratic error

For f(x)=2x−x²/4, the second derivative is f”(x)=−1/2, so M=1/2. The interval length b−a is eight.

With h=2, the bound becomes 8×4×(1/2)/12=4/3. The actual four-trapezium error is exactly 4/3.

With h=1, the bound becomes 8×1×(1/2)/12=1/3, matching the eight-trapezium error. The simple quadratic provides a transparent example of the bound in action.

An error bound is not a measurement uncertainty interval

The calculus bound controls approximation error for a known sufficiently smooth mathematical function under its assumptions.

A real measurement may also contain uncertainty in the x-positions or heights. That data uncertainty is separate and may be larger than the numerical integration error.

The Bukit Timah rounding and error-bounds lesson covers another source of uncertainty: values reported after measurement and rounding.

Trapeziums can estimate accumulation from a rate

Imagine a fictional measured flow-rate table at equally spaced times. The area under its rate–time graph estimates the total volume passing a point during that period.

If rates are in litres per minute and time is in minutes, the resulting area units are litres. The same mathematical method therefore connects geometry with rates.

This is the broader meaning of numerical integration: estimating accumulation from observations of how quickly a quantity changes.

Worked example 17: a two-interval rate model

An imaginary flow-rate table shows 2 L/min at minute zero, 4 L/min at minute five and 3 L/min at minute ten.

First trapezium gives 5(2+4)/2=15 L. Second gives 5(4+3)/2=17.5 L.

Estimated accumulated volume is 32.5 L under straight-line interpolation between the rate samples. It is not a record of actual rainfall or water consumption in Bukit Timah.

Accumulation and instantaneous rates must not be confused

A rate measured in litres per minute is not itself a volume. Multiplying that rate by an appropriate time interval produces a volume contribution.

The trapezoidal rule uses an average of two neighbouring rates over each interval. If a student adds rates without including time, the result still has rate units rather than the desired accumulated litres.

Reading units first prevents a surprisingly large class of otherwise plausible-looking errors.

Trapezoids on a graph require the correct horizontal separation

Suppose plotted points appear equally spaced in a diagram, but the x-axis labels are 0,1,3,6 rather than 0,2,4,6.

The visual gap may not reflect actual horizontal distance. The correct trapezium width comes from the numerical difference between neighbouring x-values, not pixel spacing.

A graph with compressed or inconsistent axis scales can mislead. Use the table and its units whenever numerical information is available.

Worked example 18: graph spacing is not interval width

Take sample positions x=1 and x=4, with heights two and six. Their numerical horizontal separation is three, even if the printed points are drawn only one centimetre apart.

The trapezoidal area is 3(2+6)/2=12 in the corresponding product units.

Using printed paper distance without a declared scale would calculate a different and unsupported quantity.

A trapezoidal estimate can be larger than a rectangle estimate

Suppose a graph segment has endpoint heights five and seven over width three. The trapezium estimate is 3(5+7)/2=18.

A rectangle using the smaller endpoint height five gives 15, while a rectangle using the larger height seven gives 21.

For a linear interpolation between these two endpoint heights, the trapezium area lies exactly between those two endpoint rectangles. For an unknown curve, the true area may lie elsewhere unless extra bounds are supplied.

Worked example 19: three simple ways to compare

For endpoints five and seven over width three: lower-endpoint rectangle area =3×5=15. upper-endpoint rectangle area =3×7=21. trapezium area =3×(5+7)/2=18.

The trapezium is the arithmetic average of those two rectangle totals. This is an intuitive bridge from simple area calculations to numerical integration.

What happens if one sampled height is entered incorrectly?

Our original five heights were 0,3,4,3,0. Suppose the middle height four were mistakenly recorded as fourteen.

The composite trapezoidal estimate changes from twenty to (2/2)[0+2(3+14+3)+0]=40.

A one-value transcription error doubled the estimate in this setup. The correct remedy is to verify the data source, not merely choose a smaller result because it looks aesthetically plausible.

Worked example 20: diagnose the exaggerated estimate

With erroneous table 0,3,14,3,0 and h=2, the interior sum is twenty. Twice it gives forty.

Once the middle value is confirmed as four, the interior sum becomes ten and the result returns to twenty.

Both calculations are internally consistent with their respective tables. Correctness of the input and correctness of arithmetic are separate questions.

Sampling more densely costs effort and may improve resolution

A larger number of well-positioned observations can capture curvature more accurately. It also requires additional measurement effort and may create more opportunities for inconsistent recording.

The appropriate spacing depends on how rapidly the underlying quantity changes and what accuracy is needed. For a smooth known function, mathematical error bounds guide the choice. For actual data, measurement reliability also matters.

Students should not infer that finer sampling guarantees perfect knowledge of an irregular real boundary.

A profile need not be a function in every orientation

The trapezoidal procedure as presented treats the boundary height as one y-value for each horizontal x-position.

A shape with an overhang or loop could have more than one vertical coordinate at the same x, so it may not be described by one single-valued height function without splitting it into suitable pieces.

Recognising the representation’s limits is part of choosing the correct mathematical model.

A plot with a hole requires careful interpretation

If f(x) is undefined at an interior position, the graph and the region beneath it may require additional information before a simple trapezoidal rule is used.

Numerical samples at the two sides do not automatically justify connecting through an undefined gap. The function may have a removable hole, a discontinuity or an unbounded feature.

This connects to Mathematics in Bukit Timah: algebraic fractions and domain restrictions.

Choosing the number of intervals is a modelling decision

The reference quadratic produced estimates zero with one interval, sixteen with two, twenty with four and twenty-one with eight.

The exact area is 64/3≈21.333. In this example, using narrower equal intervals improves the result.

The sequence of estimates reveals convergence toward the integral under the smooth-function assumptions. It does not mean every arbitrary sampled dataset must improve monotonically when more points are added.

Worked example 21: compare absolute errors

For the known quadratic, the two-interval estimate is 16 with error magnitude 16/3≈5.333. Four intervals give 20 with error 4/3≈1.333.

Eight intervals give 21 with error 1/3≈0.333. Each halving of step width reduced error by factor four here.

This pattern is consistent with the h² error dependence of the composite trapezoidal rule for smooth curves, though the precise constant depends on the function.

A numerical estimate is not an exact boundary survey

A professional area calculation may require controlled coordinates, survey instruments, a legally defined boundary and appropriate accuracy and documentation.

A teaching trapezoid estimate built from invented heights establishes only how the algorithm operates. It should not be used to infer ownership, buildable area, drainage conditions or property value.

This is a case where mathematical honesty improves the lesson. A number is most useful when the reader knows exactly what model produced it.

Primary foundations: average heights and trapeziums

Primary learners can begin with rectangular area and the ordinary trapezium formula. Draw two vertical sides of different lengths and a straight top connecting them.

Calculate the area by averaging the parallel sides and multiplying by the separation. Use integer widths and heights first.

Formal numerical integration need not be introduced at this stage; understanding how one trapezium is measured is the right foundation.

Lower Secondary: tables and composite shapes

Students ready for multistep area problems can add the areas of several adjacent trapeziums and check for gaps or double-counting.

They should identify which positions form each interval and how widths change. The method combines measurement, fractions, averages and careful tabulation.

A new curved-boundary story can later make clear that the sum of trapeziums is an estimate when the actual edge is not straight.

Additional Mathematics extension: integration and error

Upper-level learners can compare trapezoidal approximations with exact integrals for functions they can integrate and examine how curvature changes the sign of the error.

Simpson’s rule, derivative-based error bounds and numerical convergence are further extensions. They should be matched to the student’s current syllabus and prerequisites rather than called compulsory for every SEC Mathematics candidate.

The key learning goal is to choose a valid method, state its conditions and interpret the resulting error honestly.

Common error 1: use full sample heights as rectangle totals

A pupil may simply add 0+3+4+3+0 and call it the area. That sum equals ten but still has the wrong interpretation and lacks the horizontal widths.

The composite trapezoidal rule first forms endpoint-weighted height sums, then multiplies by half the interval width.

The corrective question is “What horizontal distance does each height apply across?” before completing the arithmetic.

Common error 2: count endpoints twice

In the equal-spacing formula, interior heights receive coefficient two, while the first and last heights appear once.

Multiplying every height by two overcounts endpoint contributions. Conversely, omitting the internal doubling undercounts shared trapezium sides.

Draw two adjacent trapeziums and point to their shared sample height. The formula becomes visually understandable.

Common error 3: invent a common h for unequal spacing

If x-values are 0,1,3,6, the gaps are 1,2,3. Their mean gap two is not a valid universal interval width for the ordinary equal-step composite formula.

Compute each trapezium with its own width instead. The given positions are mathematical data, not optional labels.

A good explanation includes the list of actual interval widths before any area sum.

Common error 4: call a signed integral geometric area

If part of a function lies below the axis, the signed integral can contain negative contributions. Geometric region area is nonnegative and may require splitting at zeros.

The f(x)=x−1 example shows how signed total zero can coexist with geometric area one.

The repair is to specify what quantity is being accumulated and whether sign has physical meaning in that model.

Common error 5: promise accuracy without a curve assumption

Five sampled heights do not determine a unique true area. The trapezoidal answer twenty is an exact sum of the invented trapezium pieces but may differ from the area of an unknown smooth boundary.

A valid conclusion should name the interpolation and, when possible, a comparison or error bound based on suitable mathematical assumptions.

More decimal places do not remove uncertainty about an unobserved curve.

A twenty-minute parent diagnostic

Minutes 1–5: Find area of one trapezium with parallel sides three and five and separation four. Answer: sixteen square units.

Minutes 6–10: Use fictional profile heights 0,3,4,3,0 at x=0,2,4,6,8. Calculate individual trapeziums 3,7,7,3 and sum twenty.

Minutes 11–15: Apply the composite formula with h=2 and confirm twenty. Explain why internal heights are doubled.

Minutes 16–20: Compare with known quadratic f(x)=x(8−x)/4, whose exact area is 64/3. The estimate is too small by 4/3. Explain which curve assumption makes that comparison possible.

A seven-day progression with retrieval

Day one revisits rectangle and trapezium areas. Day two builds two adjacent trapeziums. Day three derives the composite equal-spacing formula.

Day four works with unequal x-intervals. Day five compares estimates against a known smooth curve. Day six explores signed area and units under a rate–time graph.

Day seven gives a new table and asks for a justified approximation with the assumptions stated. A learner who can explain where the coefficient two comes from is progressing beyond memorisation.

Five practice questions and checked answers

A. A trapezium has parallel sides 2 and 6, separated by 3. Area? Answer: 3(2+6)/2=12 square units.

B. Heights are 0,3,4,3,0 at common step width two. Composite trapezoidal area? Answer: 20 square units.

C. Heights at x=0,1,3 are 0,2,4. Approximate area with individual trapeziums. Answer: 1(0+2)/2+2(2+4)/2=1+6=7.

D. For f(x)=x² on [0,2], does one trapezium overestimate or underestimate the exact integral? Answer: overestimate; four versus 8/3.

E. A rate in L/min is graphed against minutes. What units does the area have? Answer: litres.

F. Does knowing only five sampled points uniquely determine the area under every possible connecting curve? Answer: no.

Frequently asked questions

Is the trapezoidal rule an exact area formula?

It is exact for the piecewise-linear approximation and for genuinely straight graph segments. For a curved function it is generally an approximation.

Why do the middle heights have coefficient two?

Each internal height belongs to two neighbouring trapeziums, while each endpoint belongs to only one.

Can the intervals be unequal?

Yes, but calculate the trapezium over each actual width, rather than assuming the equal-spacing composite formula applies unchanged.

Is the trapezoidal estimate always too small?

No. Concave-down curves can give underestimates, while convex curves can give overestimates under appropriate conditions.

Does more sample data always guarantee greater accuracy?

Not without appropriate smoothness, measurement quality and sampling assumptions. For the known quadratic example, finer equal spacing improves the estimate.

What if the curve is below the x-axis?

The signed integral can be negative. Geometric area requires nonnegative contributions and may need splitting at the zeros.

Are these areas measured in a real Bukit Timah garden?

No. All profiles, rates and dimensions are imaginary educational models.

Is Simpson’s rule the same method?

No. It uses a different weighting pattern and conditions and can be more accurate for appropriate smooth functions.

Why can a graph area represent litres rather than square metres?

Its units come from multiplying both axes. Litres per minute multiplied by minutes gives litres.

How can a parent see genuine understanding?

The learner can build trapeziums, use the correct interval widths, verify the coefficient pattern and explain why the result is an estimate rather than a surveyed exact area.

Official context and connected Bukit Timah Mathematics

The MOE Primary Mathematics syllabus provides area and measurement foundations, while the SEAB 2027 G3 Mathematics and Additional Mathematics directory provides current subject pathways. Numerical integration, Simpson’s rule and derivative error bounds are identified here as suitable extensions rather than universal examination requirements.

For connected reasoning, read Mathematics in Bukit Timah: numerical bisection and square roots and measurement, rounding and error bounds. For a rate–time area application, see electricity use and kWh. The immutable Clementi Secondary 1 Mathematics small-group tutorial reference anchors the standard of concept-first explanations, worked checks and independent practice.

The habit to retain is to identify the baseline and units, respect actual interval widths, derive the trapezium sum, distinguish the approximation from an exact integral and state which mathematical assumptions justify the claimed accuracy.

Related Mathematics in Bukit Timah articles in this four-part batch: Graph colouring and conflict-free timetables · Binary numbers, place values and digital patterns · Standard deviation and variance. Read these connected lessons for different questions about discrete structures, positional numbers, data spread and numerical approximation.

Discover more from eduKate Singapore

Subscribe now to keep reading and get access to the full archive.

Continue reading