RMO Mock Paper 4 · IMOolympiad.com · Original practice

6 questions · 180 minutes · Written proofs

Move from finding an answer to explaining a complete argument. State your assumptions, justify the main step and account for every case.

Original independent practice. Use the suggested time for a full attempt; keep hints and solutions closed. Questions are not official RMO questions. No selection or score prediction is implied.

Question 1

Find the greatest positive integer dd that divides n5+5n3−6nn^5+5n^3-6n for every integer nn.

Hint 1

Factor the expression and inspect the moduli 3, 4 and 5.

Hint 2

The value at n=2 supplies an upper bound for d.

Worked solution 1

Factor n5+5n3−6n=n(n2−1)(n2+6)n^5+5n^3-6n=n(n^2-1)(n^2+6). One of n−1,n,n+1n-1,n,n+1 is divisible by 3. For divisibility by 4, if nn is odd then n2−1n^2-1 is divisible by 8. If nn is divisible by 4, use its own factor. If n≡2(mod4)n\equiv2\pmod4, both nn and n2+6n^2+6 are even, providing a factor 4. Modulo 5, n5≡nn^5\equiv n by checking the five residues, so the original expression is n−6n≡0n-6n\equiv0. As 3,4,5 are pairwise coprime, 60 divides every value.

At n=2n=2, the value is 32+40−12=6032+40-12=60. Any universal divisor must divide this value, hence cannot exceed 60. The required greatest divisor is 60.

Answer / conclusion: 60.

Review the idea: Divisibility · Remainders

Question 2

Find all real triples satisfying x+y+z=0x+y+z=0, x2+y2+z2=14x^2+y^2+z^2=14 and xyz=6xyz=6.

Hint 1

Find xy+yz+zx from the first two conditions.

Hint 2

Construct the monic cubic whose roots are x, y and z.

Worked solution 2

Squaring x+y+z=0x+y+z=0 gives xy+yz+zx=−7xy+yz+zx=-7. Therefore (t−x)(t−y)(t−z)=t3−7t−6=(t−3)(t+1)(t+2).(t-x)(t-y)(t-z)=t^3-7t-6=(t-3)(t+1)(t+2). The roots of the left side, with multiplicities, are exactly x,y,zx,y,z. The right side has the three distinct roots 3,−1,−23,-1,-2, so the triple is a permutation of (3,−1,−2)(3,-1,-2). Conversely this triple has sum 0, square sum 9+1+4=149+1+4=14, and product 6. All six permutations therefore work, and there are no others.

Answer / conclusion: All six permutations of (3,−1,−2).

Review the idea: Vietas formulas · Symmetric polynomials

Question 3

A convex quadrilateral ABCDABCD, with vertices in that order, is inscribed in a circle of radius RR. Its diagonals are perpendicular. Prove AB2+CD2=4R2AB^2+CD^2=4R^2. Deduce the radius when AB=5AB=5 and CD=12CD=12.

Cyclic quadrilateral with perpendicular diagonalsABCDOPOriginal construction; the proof does not rely on the drawing.

Hint 1

Use the intersecting-chords angle theorem, or compare arcs AB and CD.

Hint 2

If the half-arc angles are u and v, show u+v=90° and express the chords using sine.

Worked solution 3

Let OO be the circle’s centre and let the diagonals meet at PP. Use the four arcs AB,BC,CD,DAAB,BC,CD,DA between consecutive vertices; their measures sum to 360∘360^\circ. We first derive the needed angle relation from the inscribed-angle theorem. Since PP is on ACAC and BDBD, ∠PAB=∠CAB=12arc⁡(BC),∠PBA=∠DBA=12arc⁡(DA).\angle PAB=\angle CAB=\tfrac12\operatorname{arc}(BC),\qquad \angle PBA=\angle DBA=\tfrac12\operatorname{arc}(DA). The angle sum in triangle APBAPB therefore gives ∠APB=180∘−arc⁡(BC)+arc⁡(DA)2=arc⁡(AB)+arc⁡(CD)2.\angle APB=180^\circ-\frac{\operatorname{arc}(BC)+\operatorname{arc}(DA)}2=\frac{\operatorname{arc}(AB)+\operatorname{arc}(CD)}2.

The diagonals are perpendicular, so ∠APB=90∘\angle APB=90^\circ. Hence the two arcs AB,CDAB,CD have total measure 180∘180^\circ. Let their half-measures be u,vu,v, respectively. Both are positive and u+v=90∘u+v=90^\circ. A chord whose central angle is 2u2u has length 2Rsin⁡u2R\sin u: bisect the isosceles triangle formed by the chord and its two radii, and use sine in either right half. Applying this to the two chords gives AB=2Rsin⁡u,CD=2Rsin⁡v=2Rcos⁡u.AB=2R\sin u,\qquad CD=2R\sin v=2R\cos u. Therefore AB2+CD2=4R2(sin⁡2u+cos⁡2u)=4R2.AB^2+CD^2=4R^2(\sin^2u+\cos^2u)=4R^2. For AB=5AB=5, CD=12CD=12, this gives 4R2=25+144=1694R^2=25+144=169, so R=13/2R=13/2.

Answer / conclusion: R=13/2 for side lengths 5 and 12.

Review the idea: Circles and power of a point · Trigonometry in geometry

Question 4

Every edge joining two of seven labelled vertices is coloured red or blue. Prove that at least four triangles have all three edges the same colour, and construct a colouring with exactly four such triangles.

Hint 1

At a vertex with d red neighbours, count pairs of incident edges of different colours.

Hint 2

A non-monochromatic triangle is counted at exactly two of its vertices. For equality, try red degrees 3,3,3,3,3,3,2.

Worked solution 4

A monochromatic triangle has all three edges of the same colour. At vertex ii, let did_i be its red degree, meaning the number of red edges meeting it. Of its six incident edges, 6−di6-d_i are blue. To choose two incident edges of different colours, choose one of the did_i red edges and one of the 6−di6-d_i blue edges. This gives di(6−di)d_i(6-d_i) choices. Since di(6−di)=9−(di−3)2≤9,d_i(6-d_i)=9-(d_i-3)^2\le9, the total over seven vertices is at most 63.

Each such mixed pair determines a triangle, because the other two endpoints are joined by an edge. A triangle with two red edges and one blue edge contributes exactly two mixed pairs, one at each endpoint of its blue edge; the remaining vertex has two red edges. A triangle with two blue edges and one red edge is counted twice in the same way. A monochromatic triangle contributes no mixed pair. Thus, if TT is the number of non-monochromatic triangles, 2T=∑i=06di(6−di)≤63.2T=\sum_{i=0}^{6}d_i(6-d_i)\le63. Since TT is an integer, T≤31T\le31. There are (73)=35\binom73=35 triangles altogether, so at least 35−31=435-31=4 are monochromatic.

To attain four, label the vertices 0,1,…,60,1,\ldots,6. Colour red the seven cycle edges 01,12,23,34,45,56,6001,12,23,34,45,56,60, together with 03,15,2603,15,26. Colour every other edge blue.

Red edges: a seven-cycle and chords 03, 15, 260123456Red edges shown; every other edge is blue.
The seven-cycle and chords 03, 15, 26 are red. All other edges of the complete graph are blue.

Each cycle vertex has two red neighbours. The three additional red edges have six different endpoints, so vertices 0,1,2,3,5,60,1,2,3,5,6 have red degree 3, while vertex 4 has red degree 2. The exact counting identity now gives 2T=6(3⋅3)+2⋅4=62,T=31.2T=6(3\cdot3)+2\cdot4=62,\qquad T=31. Hence this colouring has exactly 35−31=435-31=4 monochromatic triangles, proving the bound is attainable.

Answer / conclusion: At least four, and four is attainable.

Review the idea: Counting · Combinations and binomial coefficients

Question 5

For nonnegative real numbers a,b,ca,b,c with a+b+c=1a+b+c=1, prove 4(ab+bc+ca)≤1+9abc4(ab+bc+ca)\le1+9abc and determine all equality cases.

Hint 1

Homogenise the inequality using a+b+c=1.

Hint 2

Assume a ≥ b ≥ c and group the first two terms of ∑a(a−b)(a−c)\sum a(a-b)(a-c).

Worked solution 5

Because a+b+c=1a+b+c=1, the required difference between right and left sides can be written as (a+b+c)3+9abc−4(a+b+c)(ab+bc+ca).(a+b+c)^3+9abc-4(a+b+c)(ab+bc+ca). Expanding and collecting terms shows that this is a(a−b)(a−c)+b(b−c)(b−a)+c(c−a)(c−b).a(a-b)(a-c)+b(b-c)(b-a)+c(c-a)(c-b). The original expression is symmetric, so relabelling the variables does not change it. We may therefore assume a≥b≥c≥0a\ge b\ge c\ge0.

In the second term, use b−a=−(a−b)b-a=-(a-b); in the last term, reverse both signs, so (c−a)(c−b)=(a−c)(b−c)(c-a)(c-b)=(a-c)(b-c). Factoring the first two terms then gives (a−b)(a(a−c)−b(b−c))+c(a−c)(b−c).(a-b)\bigl(a(a-c)-b(b-c)\bigr)+c(a-c)(b-c). The bracket simplifies as follows: a(a−c)−b(b−c)=a2−b2−c(a−b)=(a−b)(a+b−c).a(a-c)-b(b-c)=a^2-b^2-c(a-b)=(a-b)(a+b-c). Thus the difference is (a−b)2(a+b−c)+c(a−c)(b−c)≥0.(a-b)^2(a+b-c)+c(a-c)(b-c)\ge0. Both terms are nonnegative: a+b−c≥a≥0a+b-c\ge a\ge0, and c,a−c,b−cc,a-c,b-c are all nonnegative. This proves the inequality.

If a>ba>b, then (a−b)2>0(a-b)^2>0 and a+b−c≥a>0a+b-c\ge a>0, so the first term is positive. Equality therefore requires a=ba=b. The second term becomes c(a−c)2c(a-c)^2, which is zero precisely when c=0c=0 or c=ac=a. Together with a+b+c=1a+b+c=1, these give (a,b,c)=(1/2,1/2,0)(a,b,c)=(1/2,1/2,0) or (1/3,1/3,1/3)(1/3,1/3,1/3) in the chosen ordering. Restoring all orders, the equality cases are the permutations of (1/2,1/2,0)(1/2,1/2,0) and the equal triple (1/3,1/3,1/3)(1/3,1/3,1/3). Substitution verifies each.

Answer / conclusion: Equality at (1/3,1/3,1/3) and permutations of (1/2,1/2,0).

Review the idea: Symmetric polynomials · Sum of squares

Question 6

Let n>1n>1 be odd and suppose n∣2n+1n\mid2^n+1. Prove that 3∣n3\mid n. Also prove that infinitely many such odd integers nn exist.

Hint 1

Let p be the least prime divisor of n; every prime divisor of n is at least p.

Hint 2

The multiplicative order of 2 modulo p divides both 2n and p−1. For existence, try powers of 3.

Worked solution 6

Let pp be the smallest prime divisor of nn. Since n>1n>1 is odd, p≥3p\ge3. The assumption gives 2n≡−1(modp)2^n\equiv-1\pmod p, and squaring gives 22n≡1(modp)2^{2n}\equiv1\pmod p. By Fermat’s little theorem, 2p−1≡1(modp)2^{p-1}\equiv1\pmod p, because pp is an odd prime. Thus at least one positive exponent gives residue 1. Let dd be the least such exponent.

We need a small divisibility fact about dd. If 2m≡1(modp)2^m\equiv1\pmod p, divide mm by dd, writing m=qd+rm=qd+r with 0≤r<d0\le r<d. Then 2m=(2d)q2r≡2r(modp).2^m=(2^d)^q2^r\equiv2^r\pmod p. Hence 2r≡1(modp)2^r\equiv1\pmod p. A positive r<dr<d would contradict the minimality of dd, so r=0r=0. This proves d∣md\mid m. Applying the fact twice, we have d∣2nd\mid2n and d∣p−1d\mid p-1.

Also, gcd⁡(n,p−1)=1\gcd(n,p-1)=1: any common prime divisor would be a prime divisor of nn, so at least pp, but it would divide the smaller positive number p−1p-1, which is impossible. Therefore 2n2n and p−1p-1 can have no common odd prime factor. Since nn is odd, 2n2n contains only one factor of 2. It follows that gcd⁡(2n,p−1)≤2\gcd(2n,p-1)\le2. As dd divides both numbers, d≤2d\le2. We cannot have d=1d=1, since 2≢1(modp)2\not\equiv1\pmod p. Thus d=2d=2, so p∣22−1=3p\mid2^2-1=3, forcing p=3p=3. In particular, 3∣n3\mid n.

For infinitely many examples, we prove by induction that 3k+1∣23k+1(k≥0).3^{k+1}\mid2^{3^k}+1\qquad(k\ge0). At k=0k=0, this says 3∣33\mid3. For the induction step put A=23kA=2^{3^k}. The induction hypothesis says that A+1A+1 contains a factor 3k+13^{k+1}; it also gives A≡−1(mod3)A\equiv-1\pmod3. Consequently A2−A+1≡1+1+1≡0(mod3)A^2-A+1\equiv1+1+1\equiv0\pmod3. Factoring, 23k+1+1=A3+1=(A+1)(A2−A+1).2^{3^{k+1}}+1=A^3+1=(A+1)(A^2-A+1). The first factor is divisible by 3k+13^{k+1}, and the second supplies another factor of 3. Their product is therefore divisible by 3k+23^{k+2}, completing the induction. For every k≥1k\ge1, taking n=3kn=3^k gives an odd integer greater than 1 with n∣2n+1n\mid2^n+1. These infinitely many distinct values prove the second claim.

Answer / conclusion: 3 divides n; every positive power of 3 is an example.

Review the idea: Number theory theorems · Unique prime factorisation · Mathematical induction the first principle

After this paper

Record one idea you missed and one proof or calculation you want to improve. Work through the linked lesson, then try the next paper without hints.

Choose another RMO paper · Find a concept or theorem

Format reference: official RMO programme. Paper content is independently authored practice.