Learn the idea. Try it yourself.
Read the explanation or try 6 questions with hints. Free, untimed and no sign-in needed.
Original learning material: finite-sum practice for students exploring Singapore’s mathematical olympiad pathway. The problems teach a method through progressively harder applications; they are not copied from the official sample and are not a full mock paper.
Your goal: build and verify a useful difference, track every surviving boundary term, and use that reasoning to prove a finite-sum formula.
Before you start: subtract fractions, expand brackets, and read an inclusive integer index range. Review arithmetic and exact calculation, algebraic identities and sequences and sums when needed. The required cancellation and cube identities are derived here.
Suggested core route: 40–60 minutes for the starting checks, four examples with their nearby checks, and Questions 1–6. Take the three Stretch problems and proof extensions in a later sitting if needed. This is a flexible study estimate, not an examination duration.
Three starting checks
TS-D1. Simplify − as one fraction.
Answer and repair route
Use the common denominator 40: − = − = . The numerator is the difference of the cross-multipliers, not 1 − 1.
If needed: If fraction subtraction is uncertain, multiply each fraction by a form of 1 to give equal denominators before subtracting.
TS-D2. An indexed sum includes k = 3, 4, …, 8. How many terms does it have, and what is the last value of k + 1?
Answer and repair route
The six indices are 3, 4, 5, 6, 7 and 8. The count is 8 − 3 + 1 = 6. At the final index k = 8, the shifted expression k + 1 is 9.
If needed: Write the first and last substituted terms. The final index and the final shifted argument are different numbers.
TS-D3. Expand − .
Answer and repair route
First = + 2x + 1. Multiplying this by x + 1 gives + 3 + 3x + 1. Subtracting leaves 3 + 3x + 1.
If needed: Use the distributive law twice. The terms 3 and 3x are essential; is not + 1.
A finite sum needs an index range
We will use positive integer indices unless a question states otherwise. Every sum in this lesson is finite, and every denominator must be nonzero. A statement “for k from m through n” includes both endpoints and has terms, assuming integers .
The notation means F(m) + F(m + 1) + … + F(n). Here F(k) is simply an expression evaluated at the integer k; for instance F(k) = gives the successive cubes. The summation sign tells us what to add. It does not itself provide a method.
Why cancellation works
First consider (a − b) + (b − c) + (c − d). Combining like terms cancels b and c, leaving a − d. The same argument works for any finite chain: the negative copy of an interior expression has an equal positive copy elsewhere in the sum.
In symbols, = . Reversing the difference reverses the answer: = . Expand the first two and last two terms to check the orientation. If m = n there is one difference, and the formulas still hold.
The difficult part is often creating a difference. For a positive integer shift h, − = whenever k and k + h are nonzero. This follows because the common-denominator numerator is (k + h) − k = h. If the required numerator is different, multiply by the appropriate constant. Always verify the identity before using the cancellation.
A useful sequence: choose a possible cancelling piece, verify its difference, write the first and last terms, then cancel. A long expression is not automatically telescoping; the matching copies must actually be equal.
Worked example 1: Build a difference instead of guessing
Evaluate .
Planning hint 1 for TS-E1
The factors 2k − 1 and 2k + 1 suggest subtracting their reciprocals.
Planning hint 2 for TS-E1
Put that difference over a common denominator and compare its numerator with 2.
The two denominator factors differ by 2. Subtract their reciprocals over a common denominator: − = . This verifies the decomposition for every integer k from 3 to 10; all denominators are positive.
The sum becomes ( − ) + ( − ) + … + ( − ). Each negative reciprocal except the last is cancelled by the next positive reciprocal. Only − remains, which is .
The first index is 3, giving denominator 5. The last negative denominator is 2 × 10 + 1 = 21. Those substitutions determine the endpoints; the number of summands alone does not.
Your turn 1
Evaluate .
Hint 1 for TS-T1
Start by subtracting the reciprocals of the two factors.
Hint 2 for TS-T1
The numerator of that difference is 2, so keep the extra factor needed to obtain 4.
Full solution for TS-T1
The difference − has numerator 2. Our numerator is 4, so multiply that difference by 2. Cancellation then leaves 2( − ) = . Both denominators are nonzero throughout the stated range.
Bridge: what changes when the shift is two?
In − , the negative term at index k matches the positive term two indices later. It does not match the very next term. For indices 1 through 4, align equal denominators in columns:
| denominator | 1 | 2 | 3 | 4 | 5 | 6 |
|---|---|---|---|---|---|---|
| positive contributions | 1 | — | — | |||
| negative contributions | — | — | − | − | − | − |
| after cancellation | 1 | 0 | 0 | − | − |
For , the same arrangement gives = 1 + − − . At n = 1, the two displayed terms cancel each other and the expression still gives 1 − . Treat short ranges by substitution rather than assuming that all boundary terms are distinct.
Worked example 2: A gap of two leaves two terms at each end
Evaluate .
Planning hint 1 for TS-E2
Rewrite each summand using reciprocals whose denominators differ by two.
Planning hint 2 for TS-E2
List all positive denominator indices and all negative denominator indices; cancel their common indices.
First − = , so the required summand is half this difference. The positive reciprocals in the rewritten sum have denominators 1 through 6; the negative ones have denominators 3 through 8.
The terms with denominators 3, 4, 5 and 6 cancel. The remaining expression is ()(1 + − − ). Its value is ()( − − ) = .
Keeping only 1 − would lose both the positive and the negative . Write the two endpoint blocks before doing the fraction arithmetic.
Your turn 2
Evaluate .
Hint 1 for TS-T2
The numerator is now 2, so no factor of one half is needed.
Hint 2 for TS-T2
The first two positive denominators are 2 and 3; the final two negative denominators are 6 and 7.
Full solution for TS-T2
Here the numerator already matches − . The positive denominators run from 2 through 5 and the negative denominators from 4 through 7. Cancelling the overlap leaves + − − = .
Worked example 3: A squared denominator suggests squared reciprocals
Evaluate .
Planning hint 1 for TS-E3
The denominator is a product of two squares; try a difference of squared reciprocals.
Planning hint 2 for TS-E3
Expand − to check the numerator.
Try the difference − . Over the common denominator , its numerator is , exactly the numerator in the question. This is an identity obtained by expansion, not a pattern inferred from a few values.
After this rewrite, successive squared reciprocals cancel. The endpoints are and −, so the sum is − = . Every denominator is nonzero because k ranges over positive integers.
Your turn 3
Let n be a positive integer. If = , find n.
Hint 1 for TS-T3
Collapse the sum before trying values of n.
Hint 2 for TS-T3
The last negative term is −. Keep the positive-integer condition when solving the resulting square equation.
Full solution for TS-T3
The same identity turns the sum into 1 − . Therefore = and = 36. Since n is positive, n + 1 is positive, so n + 1 = 6 and n = 5. The negative square root is outside the stated domain. Substitution gives 1 − = .
Bridge: a polynomial can be the difference of larger expressions
The preceding examples created differences of fractions. We can also create polynomial differences. Multiplying out gives and . Expanding gives the backward version . The highest powers cancel, so a cubic difference leaves a quadratic expression; choosing the direction changes the sign of the linear term.
| expression chosen | difference used | what it can help sum |
|---|---|---|
| a linear expression | ||
| , forward | a quadratic with positive linear coefficient | |
| , backward | a quadratic with negative linear coefficient |
Choose a forward or backward cubic difference, multiply it if necessary to match the quadratic and linear coefficients, and then correct the constant. The cubes telescope, but a constant correction does not. Adding over indices m through n creates copies of that correction. Verify the resulting identity by expansion before summing it.
Worked example 4: Choose the cubic direction and keep the correction
Evaluate .
Planning hint 1 for TS-E4
Compare the signs of the linear terms in forward and backward differences of cubes.
Planning hint 2 for TS-E4
Expand − , then determine the correction needed to obtain the constant 2.
The negative linear term suggests a backward difference of cubes. Expanding gives = − 3 + 3k − 1, so . The required summand has constant 2 instead of 1. Therefore .
For k from 2 through 8, the cubic differences are ( − ) + ( − ) + … + ( − ). Their interior cubes cancel, leaving − . There are 8 − 2 + 1 = 7 indices, so the constant correction contributes seven copies of +1.
The required total is − + 7 = 512 − 1 + 7 = 518. A direct check gives 8 + 20 + 38 + 62 + 92 + 128 + 170 = 518. The choice of backward difference and the retained constant are both part of the method.
Your turn 4
Evaluate using a difference of cubes.
Hint 1 for TS-T4
First match the coefficients of and k using a multiple of the forward cubic difference.
Hint 2 for TS-T4
After doubling the cubic difference, compare its constant +2 with the required −2. The correction is applied at all six indices.
Full solution for TS-T4
The identity has the required ratio of quadratic and linear coefficients. Doubling it gives 6 + 6k + 2. To reach the required constant −2, subtract 4. Thus .
Adding over k from 1 through 6, the doubled cubic differences contribute 2( − ), while six copies of −4 contribute −24. The total is 2(343 − 1) − 24 = 684 − 24 = 660.
Pause at the first unjustified cancellation
- A missing coefficient: − has numerator 2 over the common denominator. Match the numerator before summing.
- The wrong endpoint: the final value of k + 1 is n + 1, not n. Substitute the final index into the entire expression.
- Too few surviving terms: a shift of two can leave two terms at each end. List both index ranges, then cancel their overlap.
- Cancelling unlike expressions: and are different. Only equal positive and negative copies cancel.
- Dropping a constant correction: summing −1 over an inclusive range produces minus the number of terms.
- Dividing by zero: check the original denominator before rewriting. Our rational examples use positive indices so their factors are nonzero.
- Turning a finite tail into zero: a small reciprocal at the end is still present. No infinite sum is used in this lesson.
Practise with feedback
Try six questions. Choose a starting level; the next question adjusts to your answers. Hints and worked feedback are available. Saving progress is optional and stays in this browser.
Interactive practice loads here. You can also use the complete question set below.
Nine graded practice questions
Foundation checks the rewrite and endpoints; Core combines those steps or reverses the question; Stretch asks for a complete range, a new cancelling piece or a uniqueness argument. These are levels within this lesson. Question 1 is an elementary warmup. The static questions below exactly match the marked practice bank.
TS-Q1. Elementary warmup: keep the shifted endpoint
Foundation
Find .
Hint 1 for TS-Q1
Rewrite as a difference.
Hint 2 for TS-Q1
The first positive reciprocal is ; the final negative reciprocal is .
Answer and full reasoning for TS-Q1
. Since = − , the middle terms cancel and the sum is − = . At k = 9, the negative denominator is k + 1 = 10.
Check your reasoning: Do not replace the last shifted denominator by the final index 9.
TS-Q2. Determine the missing coefficient
Foundation
For every positive integer k, = A( − ). What is A?
- 5
- 15
Hint 1 for TS-Q2
Subtract the reciprocals over their common denominator.
Hint 2 for TS-Q2
The numerator of − is 3, not 1.
Answer and full reasoning for TS-Q2
. The reciprocal difference is . The right side is therefore . All denominators are nonzero, so equality requires 3A = 5, hence A = . This constant makes the identity true for every positive integer k.
Check your reasoning: The gap between the factors determines the coefficient. Verify it rather than assuming it is 1.
TS-Q3. Repair a cancellation error
Foundation
A student evaluates but keeps only 1 − . What must be added to the student’s result to obtain the correct sum?
Hint 1 for TS-Q3
Write the four differences and look for uncancelled terms.
Hint 2 for TS-Q3
The student missed one positive term and one negative term.
Answer and full reasoning for TS-Q3
. The complete surviving expression is 1 + − − . Compared with 1 − , the missing contribution is − = . Terms with denominators 3 and 4 cancel; denominators 2 and 5 do not.
Check your reasoning: A shift of two generally leaves two terms at each end, not just one.
TS-Q4. Recover an endpoint from the sum
Core
Let n be an integer with . If = , find n.
- 9
- 18
- 19
- 20
Hint 1 for TS-Q4
The starting index is 2, so the first survivor is .
Hint 2 for TS-Q4
Set − = .
Answer and full reasoning for TS-Q4
19. Telescoping gives − = . Thus = , so n + 1 = 20 and n = 19. This lies in the required domain, and substituting it gives − = . The reciprocal equation has only this positive solution.
Check your reasoning: The final denominator is n + 1; the question asks for n.
TS-Q5. Find the missing summand
Core
From , exactly one summand with index j is removed. The remaining sum is . What is j?
- 3
- 5
- 10
- 4
Hint 1 for TS-Q5
Find the complete sum before looking for the missing term.
Hint 2 for TS-Q5
The removed value is − .
Answer and full reasoning for TS-Q5
4. The complete sum is 1 − = = . The removed value is therefore = . Hence j(j + 1) = 20. The positive integer j = 4 works. It is unique: for j ≤ 3 the product is at most 12, and for j ≥ 5 it is at least 30. Also j = 4 is within 1 through 10.
Check your reasoning: The reported total is the sum left after deletion, not the value of the deleted term.
TS-Q6. Recognise a cubic difference
Core
Evaluate .
- 485
- 486
- 343
- 511
Hint 1 for TS-Q6
Compare the summand with the expansion of − .
Hint 2 for TS-Q6
The first subtracted cube is , and the final added cube is .
Answer and full reasoning for TS-Q6
485. Expansion gives 3 + 3k + 1 = − . Adding from k = 3 through 7 cancels through , leaving − = 512 − 27 = 485. There is no additional constant correction because the summand already equals the entire cubic difference.
Check your reasoning: A cubic difference leaves the outer cubes, not the first and last index cubes in the same orientation.
TS-Q7. An inverse polynomial sum
Stretch
Let n be a positive integer. If = 440, find n.
- 9
- 10
- 11
- 12
Hint 1 for TS-Q7
Use the cubic-difference identity to express the total in terms of n.
Hint 2 for TS-Q7
Three times the sum is − 1 − n = n(n + 1)(n + 2).
Answer and full reasoning for TS-Q7
10. Let S be the required sum. The identity follows by expanding the cube. Adding it for k from 1 through n gives 3S = − 1 − n: the interior cubes cancel, and there are n copies of −1. Expanding and factoring gives 3S = + 3 + 2n = n(n + 1)(n + 2). Dividing by 3 yields S = . Since S = 440, we obtain n(n + 1)(n + 2) = 1320. Since 10 × 11 × 12 = 1320, n = 10 works. It is unique because increasing a positive integer n increases each positive factor, so the product is strictly increasing. The formula then gives 440 as required.
Check your reasoning: Finding one value is not a full uniqueness argument; explain why a larger or smaller positive n cannot give the same total.
TS-Q8. Find how many endpoints satisfy a bound
Stretch
How many positive integers n satisfy < ?
- 4
- 6
- 5
- Infinitely many
Hint 1 for TS-Q8
Each new summand is positive, so the partial sums strictly increase.
Hint 2 for TS-Q8
After cancellation the sum is − − . Compare n = 5 and n = 6 exactly.
Answer and full reasoning for TS-Q8
5. Telescoping gives = − − . For n = 1 the two middle terms cancel, so this formula also covers the smallest index. We have = < because 125 < 126, while = > because 345 > 336. Adding the next positive summand strictly increases . Consequently n = 1, 2, 3, 4, 5 work and all fail. There are exactly 5 possibilities.
Check your reasoning: Check the strict inequality using exact fractions, then use positivity to cover every smaller and larger endpoint.
TS-Q9. Choose a larger cancelling piece
Stretch
Evaluate .
Hint 1 for TS-Q9
Try subtracting the neighbouring fractions and .
Hint 2 for TS-Q9
Their common-denominator numerator is (k + 2) − k = 2.
Answer and full reasoning for TS-Q9
. The identity − = follows by putting the left side over its common denominator. The second fraction at index k is the first fraction at index k + 1. Thus the sum collapses to − = − = . All factors are positive on the given range.
Check your reasoning: A cancelling piece can itself be a product-denominator fraction; it need not be a single reciprocal.
Three optional written-proof extensions
Try one after the core route. State all index conditions, justify the identity for a general term, and account for each boundary. The checklists support self-review and are not official marking schemes.
TS-C1. Prove the shifted boundary rule, even when the blocks overlap
For positive integers n and h, prove = − . Your proof must also cover .
Hint 1 for TS-C1
Rewrite as − .
Hint 2 for TS-C1
Introduce the full sum of reciprocals from 1 through n + h, and express each of the two blocks as that full sum minus its omitted end.
Full proof for TS-C1
Since k and h are positive, neither k nor k + h is zero. A common denominator gives = − . The required left side is therefore the sum of reciprocals from 1 through n minus the sum of reciprocals from h + 1 through n + h.
Let T be the sum of all reciprocals from 1 through n + h. The first block is T minus the reciprocals from n + 1 through n + h. The second block is T minus the reciprocals from 1 through h. Subtracting these two expressions cancels T and leaves exactly the claimed right side.
This argument does not assume any order between n and h: each omitted block is a well-defined end of the same full sum. If n < h, the two boundary blocks on the right overlap, and their common terms cancel. For n = 1, the positive boundary block runs from denominator 1 through h and the negative block from denominator 2 through h + 1. Their common terms cancel, leaving 1 − 1/(h + 1) = h/(h + 1), the original one-term sum. This includes h = 1, when each block contains just one term.
Self-review checklist
- State that n and h are positive integers and all denominators are nonzero.
- Verify the rational identity over a common denominator.
- Compare both blocks with the same full reciprocal sum T.
- Explain why the proof still works when , instead of drawing disjoint blocks that need not exist.
TS-C2. Build the sum of squares from a difference of cubes
For every positive integer n, prove = using differences of cubes and the arithmetic sum 1 + 2 + … + n. Explain the constant correction.
Hint 1 for TS-C2
Expand − , then add from k = 1 through n.
Hint 2 for TS-C2
The linear terms contribute 3(1 + 2 + … + n), and the constant 1 appears n times.
Full proof for TS-C2
Let A be the sum of from k = 1 through n, and let B = 1 + 2 + … + n. Pairing B with its reverse gives n pairs of value n + 1, so 2B = n(n + 1) and B = .
Expansion gives . Sum this identity from k = 1 through n. The left side telescopes to − 1. The right side is 3A + 3B + n: there are n copies of the constant 1.
Substitute the formula for B. We obtain 3A = − 1 − − n. Expanding the cube and multiplying by 2 gives 6A = 2 + 3 + n. Factor the right side as n(2 + 3n + 1) = n(n + 1)(2n + 1), and divide by 6. This proves the formula. At n = 1 it gives 1, matching the single square.
Self-review checklist
- Derive or justify the arithmetic-sum formula used in the proof.
- Expand the cubic difference correctly.
- Explain both the endpoint cubes and the n constant terms.
- Show the final rearrangement and factorisation; check the smallest permitted n.
TS-C3. A finite sum approaches a boundary without reaching it
For each positive integer n, let = . Prove = − . Deduce that < for every permitted n and that strictly increases with n. No infinite-series result is needed.
Hint 1 for TS-C3
Subtract from and check the numerator.
Hint 2 for TS-C3
Your summand is half that difference. Keep the last product denominator, which never becomes infinite in a finite sum.
Full proof for TS-C3
The common denominator shows that − = . Hence the summand defining is half this difference. All factors are positive for .
Adding the identity cancels all the intermediate product-denominator fractions. Thus = ()( − ) = − . For n = 1, this gives − = , the original single term.
The subtracted fraction is positive for every positive integer n, so is strictly less than ; equality never occurs at a finite endpoint. Also − is the newly added summand , which is positive. Therefore strictly increases with n. These conclusions concern finite sums and require no claim about an infinite sum.
Self-review checklist
- Verify the difference identity and its factor of one half.
- Identify the first and last uncancelled fractions.
- Use positivity of the tail to prove a strict bound, not equality.
- Use the positive added term to justify strict increase.
Rebuild the cancellation independently
Without the solution, explain why the coefficient in is one half, and name all survivors when k runs from 2 through 5. Then explain why a constant correction contributes seven copies when the indices run from 2 through 8. Return later to any question that needed a hint before treating the method as secure.
Continue with other finite-sum methods, or explore consecutive integer sums for a different question: which totals admit a positive consecutive representation. Return to the foundations path or the Singapore guide for the broader route.
Singapore preparation context
The Singapore Mathematical Society’s official sample collection includes Open examples using finite polynomial sums and discusses the method of differences. This supports learning how to choose an algebraic formulation and track cancellation. The sample is undated; it does not establish the current paper’s complete topic distribution or difficulty.
Use the official Open rules and the SMO hub for the organiser’s current instructions. The multiple-choice questions here provide learning feedback, while the proof extensions develop written reasoning. All lesson exercises and explanations were authored independently; no official sample question was copied, translated or merely rescaled. Source context checked 8 October 2026.