Scribe logoScribeNotes & practice
Sign inStart free
  1. Notes
  2. ›Edexcel A-level Mathematics
  3. ›P Pure mathematics
  4. ›P9 Numerical methods

Edexcel A-level Mathematics · P · Pure mathematics

P9 Numerical methods

Locate roots by sign change, solve equations using iteration and the Newton–Raphson method, and estimate areas under curves with the trapezium rule.

  • 5 min read
  • 6 parts
  • 10 questions
  • Draft
Start a Quick 5 on this topic→Jump to the 10 questions
1/6A change of sign shows where a root lies
  1. 1A change of sign shows where a root lies
  2. 2Iteration turns an equation into a repeating process
  3. 3Newton–Raphson uses tangents to find roots quickly
  4. 4Newton–Raphson can fail if the start is poor
  5. 5The trapezium rule estimates area using strips
  6. 6Curve shape tells you whether the estimate is high or low
  7. Check your recall
  8. Practice: 10 questions
1

Part 1 of 6: A change of sign shows where a root lies

Some equations cannot be solved by rearranging, such as x3+2x−9=0x^3+2x-9=0x3+2x−9=0. In such cases we find an approximate answer instead. A numerical method builds a sequence of better and better estimates.

First you need to know where to look. A root of f(x)=0f(x)=0f(x)=0 is a value of xxx where the graph of y=f(x)y=f(x)y=f(x) meets the xxx-axis. If fff is continuous and f(a)f(a)f(a) and f(b)f(b)f(b) have opposite signs, a root lies between aaa and bbb.

Common mistake

A sign change can mislead if the graph has a break. The graph of y=1xy=\frac{1}{x}y=x1​ changes sign across x=0x=0x=0 but has no root. Two roots in one interval can also give no sign change.

Worked example

Locating a root

Show that f(x)=x3+2x−9f(x)=x^3+2x-9f(x)=x3+2x−9 has a root between 1 and 2.
  1. 1
    Evaluate at 1: f(1)=1+2−9=−6f(1)=1+2-9=-6f(1)=1+2−9=−6.
  2. 2
    Evaluate at 2: f(2)=8+4−9=3f(2)=8+4-9=3f(2)=8+4−9=3.
  3. 3
    fff is a polynomial, so it is continuous, and f(1)f(1)f(1) and f(2)f(2)f(2) have opposite signs.

Answer: There is a change of sign, so a root lies between x=1x=1x=1 and x=2x=2x=2.

Key terms · tap one for its definition

numerical method

A numerical method finds an approximate solution, usually by repeating a process that improves an estimate.

root

A root is an input ccc for which f(c)=0f(c)=0f(c)=0.

Part 1 done

2

Part 2 of 6: Iteration turns an equation into a repeating process

Once a root is located, the next job is to improve the estimate. Rearrange the equation into the form x=g(x)x=g(x)x=g(x). Choose a starting value x0x_0x0​, then repeat xn+1=g(xn)x_{n+1}=g(x_n)xn+1​=g(xn​). Each repeat is one iteration.

If the values settle towards one number, the sequence will converge to a root. This is not guaranteed. Another rearrangement of the same equation may make the values move away from the root.

Worked example

Using an iterative formula

The equation x3−x−1=0x^3-x-1=0x3−x−1=0 can be rearranged to x=x+13x=\sqrt[3]{x+1}x=3x+1​. Use xn+1=xn+13x_{n+1}=\sqrt[3]{x_n+1}xn+1​=3xn​+1​ with x0=1.5x_0=1.5x0​=1.5.
  1. 1
    x1=1.5+13=2.53≈1.3572x_1=\sqrt[3]{1.5+1}=\sqrt[3]{2.5}\approx1.3572x1​=31.5+1​=32.5​≈1.3572.
  2. 2
    x2=1.3572+13≈1.3309x_2=\sqrt[3]{1.3572+1}\approx1.3309x2​=31.3572+1​≈1.3309.
  3. 3
    x3=1.3309+13≈1.3259x_3=\sqrt[3]{1.3309+1}\approx1.3259x3​=31.3309+1​≈1.3259.

Answer: The values settle near 1.325, so the root is approximately 1.32 to 2 decimal places.

Exam tip

Keep the full calculator value between steps. Round only when you state the final answer, to the accuracy the question asks for.

Key terms · tap one for its definition

iteration

An iteration is one application of a rule such as xn+1=g(xn)x_{n+1}=g(x_n)xn+1​=g(xn​), using the previous value to produce the next.

converge

A sequence of estimates converges when its values get closer and closer to one fixed number.

Part 2 done

Quick check

One question on the part you just read

Easy

Identify the Newton–Raphson formula for the next estimate xn+1x_{n+1}xn+1​.
Show answer and explanation

Answer: B · Worked explanation

The tangent at xnx_nxn​ has gradient f′(xn)f'(x_n)f′(xn​) and meets the axis where x=xn−f(xn)f′(xn)x=x_n-\frac{f(x_n)}{f'(x_n)}x=xn​−f′(xn​)f(xn​)​. Adding the fraction instead moves the estimate in the wrong direction. Inverting the fraction gives a step of the wrong form, because the step must be a height divided by a gradient.
3

Part 3 of 6: Newton–Raphson uses tangents to find roots quickly

Simple iteration can be slow. The Newton–Raphson method usually reaches a root much faster. Start with a guess xnx_nxn​ near the root and draw the tangent to y=f(x)y=f(x)y=f(x) at that point. The tangent approximates the curve near xnx_nxn​, so a close start works best.

The tangent meets the xxx-axis at a better estimate. Its gradient is f′(xn)f'(x_n)f′(xn​), so solving for the intercept gives xn+1=xn−f(xn)f′(xn)x_{n+1}=x_n-\frac{f(x_n)}{f'(x_n)}xn+1​=xn​−f′(xn​)f(xn​)​.

Each new value becomes the next starting point. The method needs f′(x)f'(x)f′(x) to exist near the root. When it works, the values converge in only a few steps.

Figure 1. The tangent at the first estimate meets the x-axis at a closer estimate of the root.

Worked example

Newton–Raphson in practice

Use Newton–Raphson on f(x)=x3−x−1f(x)=x^3-x-1f(x)=x3−x−1 with x0=1.5x_0=1.5x0​=1.5. Find x1x_1x1​ and x2x_2x2​.
  1. 1
    Differentiate: f′(x)=3x2−1f'(x)=3x^2-1f′(x)=3x2−1.
  2. 2
    At x0=1.5x_0=1.5x0​=1.5: f(1.5)=0.875f(1.5)=0.875f(1.5)=0.875 and f′(1.5)=5.75f'(1.5)=5.75f′(1.5)=5.75.
  3. 3
    x1=1.5−0.8755.75≈1.3478x_1=1.5-\frac{0.875}{5.75}\approx1.3478x1​=1.5−5.750.875​≈1.3478.
  4. 4
    At x1x_1x1​: f(x1)≈0.1007f(x_1)\approx0.1007f(x1​)≈0.1007 and f′(x1)≈4.450f'(x_1)\approx4.450f′(x1​)≈4.450.
  5. 5
    x2=1.3478−0.10074.450≈1.3252x_2=1.3478-\frac{0.1007}{4.450}\approx1.3252x2​=1.3478−4.4500.1007​≈1.3252.

Answer: x1≈1.3478x_1\approx1.3478x1​≈1.3478 and x2≈1.3252x_2\approx1.3252x2​≈1.3252.

Key term

Newton–Raphson method

The Newton–Raphson method is an iterative method using xn+1=xn−f(xn)f′(xn)x_{n+1}=x_n-\frac{f(x_n)}{f'(x_n)}xn+1​=xn​−f′(xn​)f(xn​)​ to approach a root of f(x)=0f(x)=0f(x)=0.

Part 3 done

4

Part 4 of 6: Newton–Raphson can fail if the start is poor

Now that you can apply the formula, consider when it breaks down. A poor starting value can send the estimates far from the root. The method fails or misleads in these situations.

  • f′(xn)=0f'(x_n)=0f′(xn​)=0, so the tangent is horizontal and never meets the axis
  • xnx_nxn​ is near a stationary point, so the next value jumps far away
  • the start is nearer another root, so the method finds that one instead

If the values will not settle, choose a new starting value. Use a sign change first, so the start lies close to the root you want.

To give a root to 2 decimal places, test the rounding bounds. For f(x)=x3−x−1f(x)=x^3-x-1f(x)=x3−x−1, f(1.315)≈−0.041f(1.315)\approx-0.041f(1.315)≈−0.041 and f(1.325)≈0.001f(1.325)\approx0.001f(1.325)≈0.001. The sign change shows the root is 1.32 to 2 decimal places.

Exam tip

To prove accuracy, show a sign change between the two rounding bounds. Agreement between successive iterates is not enough.

Part 4 done

Quick check

One question on the part you just read

Easy

Identify the interval that must contain a root of f(x)=x3+2x−9f(x)=x^3+2x-9f(x)=x3+2x−9.
Show answer and explanation

Answer: A · Worked explanation

f(1)=−6f(1)=-6f(1)=−6 and f(2)=3f(2)=3f(2)=3, so the sign changes between 1 and 2 and the continuous function must cross the axis there. The interval from 2 to 3 is tempting, but f(2)=3f(2)=3f(2)=3 and f(3)=24f(3)=24f(3)=24 are both positive, so no sign change is shown.
5

Part 5 of 6: The trapezium rule estimates area using strips

Some integrals cannot be found exactly. The trapezium rule estimates the area under a curve instead. Divide the region into nnn vertical strips, each of equal strip width h=b−anh=\frac{b-a}{n}h=nb−a​.

Replace the curved top of each strip by a straight line. Each strip then becomes a trapezium. Adding their areas gives ∫aby dx≈h2[y0+yn+2(y1+y2+⋯+yn−1)]\int_a^b y\,dx\approx\frac{h}{2}\left[y_0+y_n+2(y_1+y_2+\dots+y_{n-1})\right]∫ab​ydx≈2h​[y0​+yn​+2(y1​+y2​+⋯+yn−1​)].

There are n+1n+1n+1 values of yyy for nnn strips. The first and last are counted once. Every middle value is counted twice, because it belongs to two neighbouring trapezia.

Worked example

Applying the trapezium rule

Estimate ∫131x dx\int_1^3\frac{1}{x}\,dx∫13​x1​dx using 4 strips.
  1. 1
    Strip width: h=3−14=0.5h=\frac{3-1}{4}=0.5h=43−1​=0.5.
  2. 2
    The yyy-values at x=1,1.5,2,2.5,3x=1,1.5,2,2.5,3x=1,1.5,2,2.5,3 are 1, 23, 12, 25, 131,\ \frac{2}{3},\ \frac{1}{2},\ \frac{2}{5},\ \frac{1}{3}1, 32​, 21​, 52​, 31​.
  3. 3
    Apply the rule: 0.52[1+13+2(23+12+25)]\frac{0.5}{2}\left[1+\frac{1}{3}+2\left(\frac{2}{3}+\frac{1}{2}+\frac{2}{5}\right)\right]20.5​[1+31​+2(32​+21​+52​)].
  4. 4
    Simplify: 0.25×6715=67600.25\times\frac{67}{15}=\frac{67}{60}0.25×1567​=6067​.

Answer: The estimate is 6760≈1.12\frac{67}{60}\approx1.126067​≈1.12 to 3 significant figures.

Key terms · tap one for its definition

trapezium rule

The trapezium rule approximates ∫aby dx\int_a^b y\,dx∫ab​ydx by splitting the region into strips and summing the areas of trapezia.

area

Area measures the size of a two-dimensional region.

strip width

The strip width hhh is the equal width of each strip, found from h=b−anh=\frac{b-a}{n}h=nb−a​ for nnn strips.

Part 5 done

6

Part 6 of 6: Curve shape tells you whether the estimate is high or low

An estimate is more useful when you know which way it errs. Each chord joins two points on the curve. On a curve that bends upwards, like y=1xy=\frac{1}{x}y=x1​, every chord lies above the curve.

The trapezia then include slightly too much area, so the rule overestimates. In the worked example, 6760\frac{67}{60}6067​ is above the exact value ln⁡3≈1.10\ln 3\approx1.10ln3≈1.10. On a curve that bends downwards, the chords lie below and the rule underestimates.

Effect of curve shape on the estimate

Curve bendsChords lieEstimate
upwardsabove the curvetoo large
downwardsbelow the curvetoo small
Curve bends
upwards
Chords lie
above the curve
Estimate
too large
Curve bends
downwards
Chords lie
below the curve
Estimate
too small

Using more strips makes the chords follow the curve more closely. The estimate improves, but the direction of the error stays the same.

Common mistake

Do not judge the error by whether the curve rises or falls. What matters is whether the curve bends upwards or downwards.

Note read

Check your recall[15 marks]

6 written questions. Answer each one in your head or on paper, then open the mark scheme.

  1. 1.State[2 marks]

    State the Newton–Raphson formula for xn+1x_{n+1}xn+1​ in terms of xnx_nxn​.

    Show mark schemeHide mark scheme

    Model answer

    xn+1=xn−f(xn)f′(xn)x_{n+1}=x_n-\frac{f(x_n)}{f'(x_n)}xn+1​=xn​−f′(xn​)f(xn​)​.

    • 1 mark for writing xnx_nxn​ minus a fraction
    • 1 mark for giving the fraction as f(xn)f′(xn)\frac{f(x_n)}{f'(x_n)}f′(xn​)f(xn​)​
  2. 2.Explain[2 marks]

    Explain why a change of sign in f(x)f(x)f(x) between x=ax=ax=a and x=bx=bx=b shows a root there.

    Show mark schemeHide mark scheme

    Model answer

    If fff is continuous, its graph cannot go from one side of the xxx-axis to the other without crossing it. The crossing point is a root between aaa and bbb.

    • 1 mark for stating that fff must be continuous
    • 1 mark for explaining that the graph must cross the xxx-axis between aaa and bbb
  3. 3.Calculate[3 marks]

    Calculate x1x_1x1​ when Newton–Raphson is applied to f(x)=x3+x−3f(x)=x^3+x-3f(x)=x3+x−3 with x0=1x_0=1x0​=1.

    Show mark schemeHide mark scheme

    Model answer

    f′(x)=3x2+1f'(x)=3x^2+1f′(x)=3x2+1, so f(1)=−1f(1)=-1f(1)=−1 and f′(1)=4f'(1)=4f′(1)=4. Then x1=1−−14=1.25x_1=1-\frac{-1}{4}=1.25x1​=1−4−1​=1.25.

    • 1 mark for differentiating to obtain 3x2+13x^2+13x2+1
    • 1 mark for substituting to obtain f(1)=−1f(1)=-1f(1)=−1 and f′(1)=4f'(1)=4f′(1)=4
    • 1 mark for obtaining x1=1.25x_1=1.25x1​=1.25
  4. 4.Calculate[3 marks]

    Calculate the trapezium rule estimate of ∫01x3 dx\int_0^1 x^3\,dx∫01​x3dx using 2 strips.

    Show mark schemeHide mark scheme

    Model answer

    h=0.5h=0.5h=0.5 and the yyy-values are 0, 0.125 and 1. The estimate is 0.52[0+1+2(0.125)]=0.3125\frac{0.5}{2}\left[0+1+2(0.125)\right]=0.312520.5​[0+1+2(0.125)]=0.3125.

    • 1 mark for finding h=0.5h=0.5h=0.5
    • 1 mark for finding the yyy-values 0, 0.125 and 1
    • 1 mark for substituting correctly to obtain 0.3125
  5. 5.Explain[3 marks]

    Explain why the trapezium rule overestimates ∫131x dx\int_1^3\frac{1}{x}\,dx∫13​x1​dx.

    Show mark schemeHide mark scheme

    Model answer

    The curve y=1xy=\frac{1}{x}y=x1​ bends upwards. Each chord therefore lies above the curve, so the trapezia include extra area.

    • 1 mark for stating that the curve bends upwards
    • 1 mark for stating that the chords lie above the curve
    • 1 mark for concluding that the trapezia give too much area
  6. 6.Describe[2 marks]

    Describe how to use an iterative formula xn+1=g(xn)x_{n+1}=g(x_n)xn+1​=g(xn​) to find a root.

    Show mark schemeHide mark scheme

    Model answer

    Choose a starting value x0x_0x0​, then substitute it into ggg to get x1x_1x1​, and repeat. If the values converge, the limit is a root.

    • 1 mark for choosing a starting value and repeatedly substituting into the formula
    • 1 mark for stating that converging values approach a root

Practice this topic

10 quick-fire multiple-choice questions, easy to hard. Tap an answer to see if you're right and the worked explanation. (The written questions with mark schemes are under “Check your recall” above.)

0 right0 of 10 answered

Includes the 2 quick checks in the note above.

3.

Easy[1 mark]

A trapezium rule estimate of ∫17f(x) dx\int_1^7 f(x)\,dx∫17​f(x)dx uses 3 strips. Determine the strip width hhh.
Show answer and explanation

Answer: D · Worked explanation

The strip width is h=b−an=7−13=2h=\frac{b-a}{n}=\frac{7-1}{3}=2h=nb−a​=37−1​=2. The value 3 is only the number of strips, and 6 is the whole interval length, which has not been divided by the number of strips. Always divide the interval length by the number of strips, never by the number of points.
4.

Medium·Calculate[2 marks]

Calculate x1x_1x1​ when Newton–Raphson is applied to f(x)=x2−5f(x)=x^2-5f(x)=x2−5 with x0=2x_0=2x0​=2.
Show answer and explanation

Answer: C · Worked explanation

f(2)=−1f(2)=-1f(2)=−1 and f′(x)=2xf'(x)=2xf′(x)=2x, so f′(2)=4f'(2)=4f′(2)=4. Then x1=2−−14=2.25x_1=2-\frac{-1}{4}=2.25x1​=2−4−1​=2.25. The value 1.75 comes from a sign error, subtracting the step 14\frac{1}{4}41​ instead of adding it. The value -0.25 is just the quotient ff′\frac{f}{f'}f′f​ and has not been combined with x0x_0x0​. Check the sign of f(2)f(2)f(2) before dividing.
5.

Medium·Calculate[2 marks]

Calculate the trapezium rule estimate of ∫02x2 dx\int_0^2 x^2\,dx∫02​x2dx using 2 strips.
Show answer and explanation

Answer: B · Worked explanation

Here h=1h=1h=1 and the yyy-values are 0, 1 and 4. The estimate is 12[0+4+2(1)]=3\frac{1}{2}\left[0+4+2(1)\right]=321​[0+4+2(1)]=3. The value 2.67 is the exact integral 83\frac{8}{3}38​, not the estimate. The value 6 forgets the factor 12\frac{1}{2}21​ in h2\frac{h}{2}2h​. Each middle yyy-value belongs to two neighbouring trapezia, which is why the rule doubles it.
6.

Medium[2 marks]

The iteration xn+1=xn+6x_{n+1}=\sqrt{x_n+6}xn+1​=xn​+6​ starts with x0=1x_0=1x0​=1. Calculate x2x_2x2​ to 3 significant figures.
Show answer and explanation

Answer: A · Worked explanation

x1=7≈2.6458x_1=\sqrt{7}\approx2.6458x1​=7​≈2.6458 and x2=8.6458≈2.9404x_2=\sqrt{8.6458}\approx2.9404x2​=8.6458​≈2.9404, which is 2.94. The value 2.65 is x1x_1x1​, not x2x_2x2​. The value 3.00 is the number the sequence converges towards, but x2x_2x2​ has not reached it. Each iteration uses the previous value, so keep x1x_1x1​ unrounded to stay accurate to 3 significant figures.
7.

Medium·Explain[2 marks]

Explain why the trapezium rule overestimates ∫131x dx\int_1^3\frac{1}{x}\,dx∫13​x1​dx.
Show answer and explanation

Answer: D · Worked explanation

On a curve that bends upwards, every chord lies above the curve, so the trapezia contain extra area. Doubling the middle values is part of the correct rule, since each belongs to two trapezia. It does not cause the overestimate.
8.

Hard·Identify[2 marks]

Identify a starting value for which Newton–Raphson fails to give a next estimate for f(x)=x3−3x+1f(x)=x^3-3x+1f(x)=x3−3x+1.
Show answer and explanation

Answer: C · Worked explanation

f′(x)=3x2−3f'(x)=3x^2-3f′(x)=3x2−3, so f′(1)=0f'(1)=0f′(1)=0. The tangent at x=1x=1x=1 is horizontal and never meets the axis, so the formula divides by zero. A start of 0 is acceptable, because f′(0)=−3f'(0)=-3f′(0)=−3 and the next estimate is 13\frac{1}{3}31​. In general, any start where the gradient is zero makes the formula impossible to use.
9.

Hard[2 marks]

A function has yyy-values 1, 2, 4, 5 and 6 at x=0, 0.5, 1, 1.5, 2x=0,\,0.5,\,1,\,1.5,\,2x=0,0.5,1,1.5,2. Calculate the trapezium rule estimate of ∫02y dx\int_0^2 y\,dx∫02​ydx using all 4 strips.
Show answer and explanation

Answer: B · Worked explanation

With h=0.5h=0.5h=0.5 the estimate is 0.52[1+6+2(2+4+5)]=0.25×29=7.25\frac{0.5}{2}\left[1+6+2(2+4+5)\right]=0.25\times29=7.2520.5​[1+6+2(2+4+5)]=0.25×29=7.25. The value 14.50 uses hhh instead of h2\frac{h}{2}2h​. The value 4.50 adds every yyy-value once, so the middle values are not doubled. Check the count first: 5 values mean 4 strips, and only the 3 middle values are doubled.
10.

Hard[2 marks]

Newton–Raphson is applied to f(x)=x2−x−3f(x)=x^2-x-3f(x)=x2−x−3 with x0=2x_0=2x0​=2. Calculate x2x_2x2​ to 3 decimal places.
Show answer and explanation

Answer: A · Worked explanation

f′(x)=2x−1f'(x)=2x-1f′(x)=2x−1. Then x1=2−−13=73x_1=2-\frac{-1}{3}=\frac{7}{3}x1​=2−3−1​=37​, and f(x1)=19f(x_1)=\frac{1}{9}f(x1​)=91​ with f′(x1)=113f'(x_1)=\frac{11}{3}f′(x1​)=311​, giving x2=73−133≈2.303x_2=\frac{7}{3}-\frac{1}{33}\approx2.303x2​=37​−331​≈2.303. The value 2.333 is only x1x_1x1​. The value 2.364 adds the second step instead of subtracting it. Keeping the exact fraction avoids rounding errors building up between steps.

Keep going

Next topicP10 VectorsPrevious topicP8 Integration

More in P · Pure mathematics

  1. P2Algebra and functionsNotes · 10 questions
  2. P3Coordinate geometryNotes · 10 questions
  3. P4Sequences and seriesNotes · 10 questions
  4. P5TrigonometryNotes · 10 questions
  5. P6Exponentials and logarithmsNotes · 10 questions
  6. P7DifferentiationNotes · 10 questions
  7. P8IntegrationNotes · 10 questions
  8. P9Numerical methodsNotes · 10 questions
  9. P10VectorsNotes · 10 questions

All Edexcel A-level Mathematics topics →

Draft: written by Scribe and checked automatically, not yet by a subject reviewer.

Practice questions are original, written by Scribe and checked automatically for correctness.

Topic codes follow the official numbering. Not endorsed by Pearson Edexcel or any exam board.

On this page

0/6 read

  1. 1A change of sign shows where a root lies
  2. 2Iteration turns an equation into a repeating process
  3. 3Newton–Raphson uses tangents to find roots quickly
  4. 4Newton–Raphson can fail if the start is poor
  5. 5The trapezium rule estimates area using strips
  6. 6Curve shape tells you whether the estimate is high or low
  7. Check your recall
  8. Practice: 10 questions
All coursesScribeHow it worksPricingPrivacyTerms

Notes and questions are written by Scribe in our own words. Topic codes follow the official course numbering so you can find your place. Not produced or endorsed by the IB, College Board, AQA, OCR, Pearson Edexcel or any exam board.

© 2026 Scribe

A Newton–Raphson step

y1.11.21.31.41.51.61.7-1-0.500.511.52x
  • Curve
  • Tangent at x₀
  • x₀ = 1.5 on the curve
  • x₁ ≈ 1.348