BWM Runde 1 Mock Paper 5 · IMOolympiad.com · Original practice

4 written-solution problems · Take-home proof practice; no fixed examination timer

This is independent preparation for a take-home competition. For actual entries, follow the organiser’s rules on independent work and permitted collaboration; our hints and solutions are for these original practice tasks only.

Original independent practice in English. Not an official paper or predicted selection test. Keep hints and solutions closed during your attempt.

How to review your proof

Check that you have used every condition, justified the main idea, covered all cases and stated the conclusion. A different complete proof can also be correct. This is a self-review checklist; written proofs are not automatically marked.

Question 1

A monic real polynomial PP of degree n≥2n\ge2 has distinct real roots a1<a2<⋯<ana_1<a_2<\cdots<a_n, with ai+1−ai>1a_{i+1}-a_i>1 for each ii. Put Q(x)=P(x+1)−P(x)Q(x)=P(x+1)-P(x). Prove that QQ has exactly one root in each interval

(ai,ai+1−1)(i=1,…,n−1),\begin{gathered}(a_i,a_{i+1}-1)\\(i=1,\ldots,n-1),\end{gathered}

that each of these roots has multiplicity 1, and that QQ has no other roots, real or complex.

Hint 1

The leading terms of P(x+1)P(x+1) and P(x)P(x) cancel. What is the degree of QQ?

Hint 2

Compare the signs of Q(ai)=P(ai+1)Q(a_i)=P(a_i+1) and Q(ai+1−1)=−P(ai+1−1)Q(a_{i+1}-1)=-P(a_{i+1}-1).

Worked solution 1

Since PP is monic of degree nn, expanding its leading term gives (x+1)n−xn=nxn−1+⋯(x+1)^n-x^n=nx^{n-1}+\cdots. Terms of lower degree in PP contribute differences of degree at most n−2n-2. Thus QQ has degree exactly n−1n-1.

Fix ii. The gap assumption places both ai+1a_i+1 and ai+1−1a_{i+1}-1 strictly between aia_i and ai+1a_{i+1}. In that interval, the factorisation

P(x)=∏j=1n(x−aj)P(x)=\prod_{j=1}^{n}(x-a_j)

has exactly n−in-i negative factors. Its sign is therefore constant and nonzero there. Consequently Q(ai)=P(ai+1)Q(a_i)=P(a_i+1) and Q(ai+1−1)=−P(ai+1−1)Q(a_{i+1}-1)=-P(a_{i+1}-1) have opposite signs.

A polynomial is continuous: as its argument changes without a jump, so does its value. The intermediate value theorem therefore gives at least one zero between these two endpoints. The n−1n-1 intervals are disjoint, so we have n−1n-1 distinct real roots of QQ.

A nonzero polynomial of degree n−1n-1 cannot have more than n−1n-1 roots counted with multiplicity: each root supplies a linear factor, lowering the remaining degree by 1. These already account for its entire degree. Hence there is exactly one in each interval, each is simple, and there are no others.

Conclusion: Exactly one simple root lies in each specified interval; these are all the roots of QQ.

Question 2

Find all pairs of positive integers (x,y)(x,y) satisfying

x3+y3=2(x2+xy+y2).x^3+y^3=2(x^2+xy+y^2).

Hint 1

Put s=x+ys=x+y and p=xyp=xy, then use x3+y3=s3−3spx^3+y^3=s^3-3sp.

Hint 2

You should obtain p(3s−2)=s2(s−2)p(3s-2)=s^2(s-2). Combine this with p≤s2/4p\le s^2/4 to bound ss.

Worked solution 2

The equation is symmetric, so use the sum s=x+ys=x+y and product p=xyp=xy. Both are positive integers, s≥2s\ge2, and the equation becomes

s3−3sp=2(s2−p),p(3s−2)=s2(s−2).\begin{gathered}s^3-3sp=2(s^2-p),\\ p(3s-2)=s^2(s-2).\end{gathered}

The left side is positive, so s>2s>2. The inequality (x−y)2≥0(x-y)^2\ge0 gives 4p≤s24p\le s^2. Substituting the required value of pp and cancelling positive quantities gives

s2(s−2)3s−2≤s24,4(s−2)≤3s−2,s≤6.\begin{gathered}\frac{s^2(s-2)}{3s-2}\le\frac{s^2}{4},\\4(s-2)\le3s-2,\\ s\le6.\end{gathered}

Only s=3,4,5,6s=3,4,5,6 remain. They give p=9/7,16/5,75/13,9p=9/7,16/5,75/13,9, respectively. The first three are not integers. For s=6,p=9s=6,p=9,

(x−y)2=s2−4p=0,(x-y)^2=s^2-4p=0,

so x=y=3x=y=3. Direct substitution gives 27+27=2(9+9+9)27+27=2(9+9+9), proving that this pair works and that there are no other positive integer solutions.

Conclusion: (x,y)=(3,3)(x,y)=(3,3) only.

Question 3

Let nn be a positive integer with gcd⁡(n,6)=1\gcd(n,6)=1. Construct an n×nn\times n array whose entries belong to {0,1,…,n−1}\{0,1,\ldots,n-1\}, so that each row, each column, the main diagonal, and the other diagonal each contain every number in this set exactly once. Prove that your construction works.

Hint 1

Number rows and columns 0,1,…,n−10,1,\ldots,n-1, and try a linear expression in the two indices modulo nn.

Hint 2

Use the remainder of i+2ji+2j in cell (i,j)(i,j). On the two diagonals this becomes 3i3i and −i−2-i-2 modulo nn.

Worked solution 3

Number rows and columns from 0 to n−1n-1. In cell (i,j)(i,j), place the unique remainder of i+2ji+2j on division by nn.

We use a basic fact: if gcd⁡(c,n)=1\gcd(c,n)=1, then the remainders of ct+dct+d, for t=0,…,n−1t=0,\ldots,n-1, are all different. Indeed, equal remainders at t1,t2t_1,t_2 imply n∣c(t1−t2)n\mid c(t_1-t_2), hence n∣t1−t2n\mid t_1-t_2. The difference has absolute value less than nn, so it must be zero. Since there are nn different remainders, all possible remainders occur.

  • In a fixed row ii, the entries are 2j+i2j+i; the coefficient 2 is coprime to nn.
  • In a fixed column jj, they are i+2ji+2j; the coefficient 1 is coprime to nn.
  • On the main diagonal, j=ij=i, so the entries are 3i3i; the coefficient 3 is coprime to nn.
  • On the other diagonal, j=n−1−ij=n-1-i, so the entries are congruent to −i−2-i-2; the coefficient −1-1 is coprime to nn.

The assumed gcd⁡(n,6)=1\gcd(n,6)=1 supplies both needed coprimality conditions. Thus every required line contains each symbol exactly once. For n=1n=1, the construction is the single entry 0 and also works.

Conclusion: Place i+2j(modn)i+2j\pmod n in cell (i,j)(i,j), using representatives 0,…,n−10,\ldots,n-1.

Question 4

In triangle ABCABC, suppose AB>ACAB>AC. The internal bisector of ∠BAC\angle BAC meets BCBC at DD; the external bisector meets line BCBC at EE, beyond CC. If MM is the midpoint of BCBC, prove that

MD⋅ME=MB2.MD\cdot ME=MB^2.Internal and external angle bisectors, with E beyond CABCDEMOriginal construction. The proof does not rely on the drawing.

Hint 1

Write BC=aBC=a, AC=bAC=b, AB=cAB=c, with c>bc>b. Use both angle-bisector ratios.

Hint 2

Obtain BD=ac/(b+c)BD=ac/(b+c) and BE=ac/(c−b)BE=ac/(c-b), then subtract BM=a/2BM=a/2.

Worked solution 4

Let BC=aBC=a, AC=bAC=b, and AB=c>bAB=c>b. The internal angle-bisector theorem gives BD/DC=c/bBD/DC=c/b, hence

BD=acb+c.BD=\frac{ac}{b+c}.

For completeness, the external ratio is BE/CE=c/bBE/CE=c/b. One way to see it is to compare the areas of triangles ABEABE and ACEACE. Using bases on line BCBC, the ratio is BE/CEBE/CE, since they share an altitude from AA. Using the common side AEAE, the ratio is ABsin⁡∠BAE/(ACsin⁡∠CAE)=c/bAB\sin\angle BAE/(AC\sin\angle CAE)=c/b: the two indicated angles are supplementary for an external bisector and therefore have equal sines. As EE is beyond CC, CE=BE−aCE=BE-a, giving

BEBE−a=cb,BE=acc−b.\begin{gathered}\frac{BE}{BE-a}=\frac cb,\\ BE=\frac{ac}{c-b}.\end{gathered}

Since c>bc>b, we have BD>a/2BD>a/2, so the order is B,M,D,C,EB,M,D,C,E. It follows that

MD=acb+c−a2=a(c−b)2(b+c),ME=acc−b−a2=a(c+b)2(c−b).\begin{gathered}MD=\frac{ac}{b+c}-\frac a2=\frac{a(c-b)}{2(b+c)},\\ ME=\frac{ac}{c-b}-\frac a2=\frac{a(c+b)}{2(c-b)}.\end{gathered}

Multiplying cancels the side-length ratios and leaves MD⋅ME=a2/4=MB2MD\cdot ME=a^2/4=MB^2, as required.

Conclusion: MD⋅ME=BC2/4=MB2MD\cdot ME=BC^2/4=MB^2.

After this paper

Choose one gap in your proof to repair, study the linked idea, and write a complete solution again before the next mock.

Choose another paper · Check your German selection route

Format reference: official organiser information. Questions and explanations are independent practice material.