IMO 1963 Problem 3
The first issue is whether the hypothesis “consecutive sides satisfy $a_1 \ge a_2 \ge \cdots \ge a_n$” is purely linear or implicitly cyclic.
Proposed by: -
Verified: no
Verdicts: FAIL + UNKNOWN
Solve time: 14m22s
Problem
In an $n$-gon all of whose interior angles are equal, the lengths of consecutive sides satisfy the relation
$a_1\ge a_2\ge \cdots \ge a_n$.
Prove that $a_1=a_2=\cdots = a_n$.
Exploration
The first issue is whether the hypothesis “consecutive sides satisfy $a_1 \ge a_2 \ge \cdots \ge a_n$” is purely linear or implicitly cyclic. In a polygon, “consecutive sides” are understood cyclically, so the pair $(a_n, a_1)$ is also a pair of consecutive sides. This immediately suggests the additional inequality $a_n \ge a_1$, which closes the chain.
Testing small cases confirms this interpretation. For $n=3$, the conditions read $a_1 \ge a_2 \ge a_3 \ge a_1$, which forces equality. For $n=4$, we similarly obtain $a_1 \ge a_2 \ge a_3 \ge a_4 \ge a_1$, again forcing equality. The same cyclic closure persists for all $n \ge 3$.
No trigonometric or geometric structure beyond cyclic ordering is needed, and attempts to use Fourier expansions are unnecessary once the cyclic meaning of “consecutive” is correctly enforced.
Problem Understanding
An $n$-gon has equal interior angles, but the only relevant hypothesis is that its side lengths in cyclic order satisfy a nonincreasing chain $a_1 \ge a_2 \ge \cdots \ge a_n$, where each comparison is between consecutive sides of the polygon.
Since the sides are arranged in a closed cycle, the last side is also consecutive to the first, so the inequality system includes the additional constraint $a_n \ge a_1$. The goal is to deduce that all side lengths are equal.
Key Observations
The ordering hypothesis forms a closed loop of inequalities along the polygon. Any cyclic chain of weak inequalities in a single direction forces all terms to coincide, since a strict inequality at any position would propagate around the cycle and contradict its return to the starting point.
No geometric angle information is needed; the conclusion is purely combinatorial once cyclic adjacency is correctly interpreted.
Solution
The hypothesis that consecutive sides satisfy
$a_1 \ge a_2 \ge \cdots \ge a_n$
is understood in cyclic order around the polygon, so each inequality compares adjacent sides and the last side is consecutive to the first. Therefore the full set of inequalities includes
$a_1 \ge a_2, \quad a_2 \ge a_3, \quad \ldots, \quad a_{n-1} \ge a_n, \quad a_n \ge a_1.$
Combining the first and last inequalities gives
$a_1 \ge a_2 \ge \cdots \ge a_n \ge a_1.$
This chain implies $a_1 \ge a_1$, with every intermediate term bounded above and below by $a_1$. If any strict inequality occurred at some index $k$, then following the chain would produce $a_1 > a_1$, which is impossible. Hence no strict inequality can occur anywhere in the cycle.
Therefore all inequalities are equalities, so
$a_1 = a_2 = \cdots = a_n.$
This completes the proof. ∎
Verification of Key Steps
The only structural point requiring validation is the cyclic interpretation of “consecutive sides.” In any polygon, side $a_n$ is consecutive to side $a_1$, so the monotonic chain must include the inequality $a_n \ge a_1$. This closes the system into a cycle.
Once the cycle of inequalities is established, transitivity forces every term to equal every other term, since any strict decrease would propagate around the entire cycle and contradict the final inequality returning to the starting point.
No hidden assumptions about angles, convexity, or coordinate representations are used, so no sign or analytic issues arise.
Alternative Approaches
A geometric reformulation uses the fact that a cyclic weakly monotone sequence on a closed loop cannot contain a strict extremum without violating adjacency consistency, yielding the same conclusion without explicitly writing the final inequality $a_n \ge a_1$.