MathLabs

Problem 1

Let n,k≥2n, k \ge 2 be positive integers and let a1,a2,…,aka_1, a_2, \dots, a_k be distinct integers in the set {1,2,…,n}\{1, 2, \dots, n\} such that nn divides ai(ai+1−1)a_i(a_{i+1} - 1) for i=1,2,…,k−1i = 1, 2, \dots, k - 1. Prove that nn does not divide ak(a1−1)a_k(a_1 - 1).
Step 5 of 5: Combine via CRT to reach the contradiction
In plain words

Agreement modulo every prime power factor of n means agreement modulo n itself, but the aia_i were chosen to be distinct residues mod n.

a1≡a2≡⋯≡ak(modn)a_1 \equiv a_2 \equiv \dots \equiv a_k \pmod n
Detailed analysis

In either case, every aia_i is congruent to the same constant (00 or 11) modulo qq. Since q=peq=p^e was an arbitrary prime power exactly dividing nn, the Chinese Remainder Theorem gives a1≡a2≡⋯≡ak(modn)a_1\equiv a_2\equiv\dots\equiv a_k\pmod n. But a1,…,aka_1,\dots,a_k are distinct elements of {1,…,n}\{1,\dots,n\}, a complete residue system mod nn, so they cannot all be congruent modulo nn (as k≥2k\ge2). This contradiction shows the assumption was false, so nn does not divide ak(a1−1)a_k(a_1-1). ■\blacksquare