The paper concerns a special property of binomial coefficients, which are the numbers that count how many ways you can choose k items from a collection of n items. A binomial coefficient is called "good" if every prime number that divides it is larger than k. Intuitively, this means the binomial coefficient has no small prime factors relative to its own index k. The "deficiency" of a good binomial coefficient measures, in a precise sense, how close it comes to satisfying an even stricter condition: it counts how many integers in a specific range near n are divisible only by primes up to k. A deficiency of zero would be the ideal case, and a deficiency of one means there is exactly one such integer causing a small obstruction.
The central question addressed here was posed decades ago by the mathematicians Paul Erdos, Lacampagne, and Selfridge, who asked whether there are infinitely many good binomial coefficients with deficiency exactly one. This sits in a broader research tradition of understanding how prime factors are distributed in binomial coefficients, a topic with deep connections to number theory and combinatorics. The existence of even a single infinite family of such objects was unresolved.
The paper provides a short, constructive proof that the answer is yes: there are infinitely many such binomial coefficients. Constructive here means the proof does not just argue abstractly that they must exist but actually builds an explicit infinite collection of examples. The approach is described as brief and direct, which is notable because questions in this area can often require heavy machinery. The result closes a long-standing open problem by confirming that deficiency one is not a rare accident but a property that occurs infinitely often.