We analyze the last digits of more than 3.4 million twin prime pairs in bases 6, 12, 18, 30, and 10. The observed digit transitions coincide exactly with those predicted by the condition that both primes must be coprime to the base. Moreover, the frequencies of the admissible transitions are nearly identical, providing empirical support for the expected equidistribution of twin primes among valid residue classes. This experiment provides a simple numerical illustration of the modular structure underlying twin primes and highlights the role of wheel factorization in prime distributions. A key observation is that the counts for base 10 and base 30 are identical, a direct consequence of the Chinese Remainder Theorem.
Victor Eduardo Morales Cordoba (2026) studied this question.