EGMO 2013 Problem 1: solution, and how to spot the hidden centroid
28 October 2026
EGMO 2013 Problem 1 is the kind of geometry question that looks like it wants a page of algebra and in fact wants one observation. Here is the statement, a full solution, a second solution you can use as a check, and the part most write-ups skip: what in the wording should have pointed you at the answer.
The signal in the statement
Read the two constructions as facts about ratios, not as lengths. says that is the midpoint of . says that sits on segment with .
A midpoint and a split on the same segment is a pattern worth knowing on sight. In triangle , the segment from to the midpoint of the opposite side is a median. A point two thirds of the way along a median, measured from the vertex, is the centroid. So is the centroid of triangle .
That is the whole problem. The triangle in the statement is a distraction; the triangle that matters is , which the statement never names. Whenever a question hands you a midpoint and a factor of two on segments that meet, try drawing the bigger triangle they belong to and ask whether a centroid is hiding in it.
Solution
- Since and is on line beyond , the point is the midpoint of . So is the median of triangle from .
- Since is on line beyond with , the point lies on that median with . The medians of a triangle meet at the point two thirds of the way from each vertex, so is the centroid of triangle .
- Let be the midpoint of . The median from also passes through the centroid, so , , are collinear, and .
- Now use the hypothesis. From we get .
- So is at distance from the midpoint of , which puts on the circle with diameter . The angle in a semicircle is a right angle, so .
- Finally , , are collinear with between and , so . The triangle is right-angled at .
Step 5 is worth a line of justification in a written proof rather than a bare citation. Triangles and are isosceles with apex , so and . The angles of triangle sum to , and is the sum of the two angles at , so .
A second solution, by vectors
When a synthetic idea will not come, this problem also falls to a short computation, and it is a good habit to use one to check the other. Put at the origin and write and for the positions of and .
- is the midpoint of , so .
- is on the far side of from with , so .
Then and . Expanding both,
So exactly when , which says . This proves the problem and, as a bonus, its converse: in a triangle right-angled at the two lengths are always equal, and in any other triangle they never are.
Where marks go missing
- Asserting a collinearity without a reason. That , , lie on one line is true because is the centroid. Say so; the centroid is the entire argument.
- Mixing up the two angles at the vertex. The circle gives , but the question is about . One sentence about , , being collinear closes the gap.
- Citing "angle in a semicircle" for a point you only know is at the right distance. That is the converse of the usual statement. It is true, but justify it with the two isosceles triangles as above.
- Proving the converse instead. Showing that a right angle at forces is the easy direction and does not answer the question.
Practise the idea
The centroid fact this rests on, that medians meet two thirds of the way along, is in the problem archive as the medians are concurrent and cut each other ; prove it once yourself and you will recognise the configuration faster. The finishing move is a piece of angle chasing, and the techniques index lists the other methods that turn up in olympiad geometry. The daily problem posts a new one every day, no account needed.
Related guides
- EGMO 2022 Problem 1: solution, and the incentre hiding in itA full solution to EGMO 2022 Problem 1: why the orthocentre is the incentre of triangle BSC, and why the circumcentre lands on the same bisector.
- Cyclic quadrilateral tricks for olympiad geometryThe cyclic quadrilateral is rarely drawn for you. Five ways to prove four points concyclic, where hidden ones sit, and what each one hands you back.
- Olympiad geometry without coordinates: where to startCoordinates are what you reach for when you do not know the synthetic first move. Translate the goal, hunt the cyclic quads, and let power of a point do the lengths.
- Power of a point: when to use it, and when not toPower of a point is a recognition problem, not a formula problem. Five tells that call for it, one that does not, and how to find the circle nobody drew.
Train it on Lemma
Every technique here is taught explicitly, in order. 128 lessons, 1237 curated problems and unlimited generated practice at six difficulties. Free to start, no card, and the paid plan opens with 3 free days.
Find your level