24 Philippine Mathematical Olympiad National Stage (Day 1) 18 March 2022
24 Philippine Mathematical Olympiad National Stage (Day 1) 18 March 2022
24 Philippine Mathematical Olympiad National Stage (Day 1) 18 March 2022
2. The PMO Magician has a special party game. There are n chairs, labelled 1 to n. There are
n sheets of paper, labelled 1 to n.
On each chair, she attaches exactly one sheet whose number does not match the number
on the chair.
She then asks n party guests to sit on the chairs so that each chair has exactly one
occupant.
Whenever she claps her hands, each guest looks at the number on the sheet attached to
their current chair, and moves to the chair labelled with that number.
Show that if 1 < m ≤ n, where m is not a prime power, it is always possible for the PMO
Magician to choose which sheet to attach to each chair so that everyone returns to their original
seats after exactly m claps.
3. Call a lattice point visible if the line segment connecting the point and the origin does not pass
through another lattice point. Given a positive integer k, denote by Sk the set of all visible
lattice points (x, y) such that x2 + y 2 = k 2 . Let D denote the set of all positive divisors of
2021 · 2025. Compute the sum
X
|Sd |.
d∈D
Here, a lattice point is a point (x, y) on the plane where both x and y are integers, and |A|
denotes the number of elements of the set A.
4. Let △ABC have incenter I and centroid G. Suppose that PA is the foot of the perpendicular
from C to the exterior angle bisector of B, and QA is the foot of the perpendicular from B to
the exterior angle bisector of C. Define PB , PC , QB , and QC similarly. Show that PA , PB ,
PC , QA , QB , and QC lie on a circle whose center is on line IG.
24th Philippine Mathematical Olympiad
National Stage (Day 2)
19 March 2022
5. Find all positive integers n for which there exists a set of exactly n distinct
positive integers, none of which exceed n2 , whose reciprocals add up to 1.
Solution. We claim that the only solutions are f (x) = 0 and f (x) = x. It is easy to see
that both satisfy the functional inequality. We now show that no other functions satisfy
the given condition.
Letting a = b = c = d yields 2f (0)2 ≤ 0 which implies f (0) = 0. Now, suppose that
f ̸≡ 0. Then we can find p ∈ R such that f (p) ̸= 0. For a ∈ R, let b = a, c = 0, and
d = a − p. We then have
f (p)f (a) ≤ af (p).
Case 1: Suppose f (p) > 0. Then f (a) ≤ a for all a ∈ R. In particular f (−1) < 0. In
the original functional equation, letting b = a, c = 0, and d = a + 1 yields f (−1)f (a) ≤
af (−1) for all a ∈ R. Hence, f (a) ≥ a for all a ∈ R. Thus for this case, f (x) = x for all
x ∈ R.
Case 2: Suppose f (p) < 0. Then f (a) ≥ a for all a ∈ R. In particular f (1) > 0. In the
original functional equation, letting b = a, c = 0, and d = a − 1 yields f (1)f (a) ≤ af (1)
for all a ∈ R. Hence, f (a) ≤ a for all a ∈ R. Thus for this case, f (x) = x for all x ∈ R.
2
2. The PMO Magician has a special party game. There are n chairs, labelled 1 to n. There
are n sheets of paper, labelled 1 to n.
On each chair, she attaches exactly one sheet whose number does not match the
number on the chair.
She then asks n party guests to sit on the chairs so that each chair has exactly one
occupant.
Whenever she claps her hands, each guest looks at the number on the sheet attached
to their current chair, and moves to the chair labelled with that number.
Show that if 1 < m ≤ n, where m is not a prime power, it is always possible for the
PMO Magician to choose which sheet to attach to each chair so that everyone returns
to their original seats after exactly m claps.
3. Call a lattice point visible if the line segment connecting the point and the origin does
not pass through another lattice point. Given a positive integer k, denote by Sk the set
of all visible lattice points (x, y) such that x2 + y 2 = k 2 . Let D denote the set of all
positive divisors of 2021 · 2025. Compute the sum
X
|Sd |.
d∈D
Here, a lattice point is a point (x, y) on the plane where both x and y are integers, and
|A| denotes the number of elements of the set A.
Solution. We claim that the required sum is 20.
Let Tk denote the set of all lattice points in the circle x2 + y 2 = k 2 . We claim that
P
d|k |Sd | = |Tk |. Indeed, given a point (x, y) in Tk , let g = gcd(x, y). Then x/g, y/g are
2 2 2
necessarily coprime, and S hence (x/g, y/g) visible, and (x/g) + (y/g) = (k/g) . This
implies (x/g, y/g) ∈ d|k Sd . Next, note that the Sd ’s are necessarily disjoint. Now
if (x′ , y ′ ) is a visible lattice point in Sd where d|k, then we can write k = gd so S that
′ ′
(x, y) = (gx , gy ) is a lattice point in Tk . This establishes a bijection between d|k Sd
and Tk , and since the Sd ’s are disjoint, the claim follows.
From the claim, it suffices to find the number of lattice points in the circle x2 + y 2 =
(2021 · 2025)2 . This is equivalent to
x2 + y 2 = 38 · 54 · 432 · 472 .
Now it is well-known that if x2 + y 2 ≡ 0 (mod p) where p ≡ 3 (mod 4) is a prime,
then x ≡ y ≡ 0 (mod p). Thus, we must also have x, y ≡ 0 (mod 34 · 43 · 47). It then
follows that the number of lattice points is the same as the number of lattice points in
x2 + y 2 = 252 .
If x = 0 or y = 0, there are 4 solutions. Otherwise, assume WLOG that they are
both positive. Now it is well-known that all solutions to x2 + y 2 = z 2 are in the form
x = g(m2 − n2 ), y = 2gmn, and z = g(m2 + n2 ), where m > n are coprime positive
integers, and g is a positive integer. Thus, we want g(m2 + n2 ) = 25. Note that g|25, so
g = 1, 5, 25.
If g = 25, then m2 + n2 = 1, so n = 0, contradiction. If g = 5, then m2 + n2 = 5,
which yields m = 2 and n = 1 and thus g(m2 − n2 ) = 15 and 2gmn = 20, so (x, y) =
(15, 20), (20, 15). If g = 1, then we get m2 + n2 = 25, from which we obtain m = 4 and
n = 3. It then follows that (x, y) = (24, 7), (7, 24), and so we have 2 solutions when x, y
are both positive. This implies that there are 4 · 4 = 16 solutions when x, y are nonzero,
and so there are 4 + 16 = 20 solutions in total. 2
4. Let △ABC have incenter I and centroid G. Suppose that PA is the foot of the perpen-
dicular from C to the exterior angle bisector of B, and QA is the foot of the perpendicular
from B to the exterior angle bisector of C. Define PB , PC , QB , and QC similarly. Show
that PA , PB , PC , QA , QB , and QC lie on a circle whose center is on line IG.
5. Find all positive integers n for which there exists a set of exactly n distinct positive
integers, none of which exceed n2 , whose reciprocals add up to 1.
Solution. The answer is all n ̸= 2. For n = 1, the set {1} works. For n = 2, no set exists,
simply because the sum of reciprocals of two distinct integers cannot be equal to 1. For
n = 3, take {2, 3, 6}.
For n > 3, the identity
1 1 1 1 1
= + + + ··· +
k k + r k(k + 1) (k + 1)(k + 2) (k + r − 1)(k + r)
allows us to extend a sum of t terms to one of exactly t + r terms. Taking k = 3 and
r = n − 3 allows us to turn the sum 1 = 1/2 + 1/3 + 1/6 to the n-term sum
1 1 1 1 1 1
1= + + + + + ··· + .
2 6 n 3·4 4·5 (n − 1)n
This construction works provided that n ̸= k(k +1) for any k. Otherwise, we have n ≥ 6,
and instead we apply the above to the sum 1 = 1/2 + 1/3 + 1/10 + 1/15, taking k = 3,
r = n − 4 to yield
1 1 1 1 1 1 1
1= + + + + + + ··· + .
2 10 15 n − 1 3 · 4 4 · 5 (n − 2)(n − 1)
The above then works, because if n = k(k + 1) for some k, then n − 1 ̸= 2, 10, 15, and
n − 1 ̸= m(m + 1) for any m by parity, since n − 1 is odd and m(m + 1) is always even.
This construction is not unique; there are other similar ones. 2
6. In △ABC, let D be the point on side BC such that AB + BD = DC + CA. The line
AD intersects the circumcircle of △ABC again at point X ̸= A. Prove that one of the
common tangents of the circumcircles of △BDX and △CDX is parallel to BC.
Solution. Refer to the figure shown below:
V E = BV sin ∠V BE
BX sin ∠BXV ∠BXD
= sin
sin ∠BDX 2
∠BXA
BX sin 2 C
= sin
sin ∠BDX 2
BX C
= sin2 .
sin ∠BDX 2
CX
Similarly, we can prove that W F = sin ∠CDX sin2 B2 . Thus, to prove the claim, it suffices
BX sin2 B2
to prove that = . This is because
CX sin2 C2
2 ≥ (1 + a4 )(2 − a2 )
(a4 + 1)(a2 − 2) + 2 ≥ 0
a6 − 2a4 + a2 − 2 + 2 ≥ 0
a6 − 2a4 + a2 ≥ 0
a2 (a4 − 2a2 + 1) ≥ 0
a2 (a2 − 1)2 ≥ 0
which is true.
Thus
1 2 − a2
≥
1 + a4 2
bc 2bc − a2 bc
≥ .
1 + a4 2
Similarly
ca 2ca − ab2 c
≥ , and
1 + b4 2
ab 2ab − abc2
≥ .
1 + c4 2
Adding these up yields
bc ca ab a2 bc + ab2 c + abc2
+ + ≥ (bc + ca + ab) − .
1 + a4 1 + b 4 1 + c 4 2
Observe that
(ab + bc + ca)2 ≥ 3((ab)(bc) + (bc)(ca) + (ca)(ab))
9 ≥ 3(a2 bc + ab2 c + abc2 )
3 ≥ a2 bc + ab2 c + abc2 .
Thus,
a2 bc + ab2 c + abc2 3 3
(bc + ca + ab) − ≥3− = .
2 2 2
Therefore,
bc ca ab 3
4
+ 4
+ 4
≥ .
1+a 1+b 1+c 2
2