The paper studies a basic question in number theory: when you take a sequence of numbers and reduce them modulo a prime, how many distinct remainders do you get? This is called the "value set" size. The sequences in question are not random but highly structured, such as products of consecutive integers (factorials), their generalizations like Gaussian q-factorials, and combinatorial quantities like Catalan numbers and binomial coefficients. The central challenge is proving that these value sets must be large, meaning the sequences cannot "collapse" onto just a few residues mod a prime.
The authors develop two general principles for establishing such lower bounds. The first applies to sequences where consecutive terms are related by simple scaling and shifting operations, and it guarantees that the value set grows at least as fast as the number of terms raised to the power 8/15. This is proved using classical geometric ideas about how many times lines and points can meet in a grid-like configuration. The second principle applies to sequences where the ratio between consecutive terms evolves according to a Mobius transformation, a fractional linear map of the form (ax+b)/(cx+d). Here the value set grows faster than the square root of the number of terms, which is a meaningful improvement. This part of the proof draws on a deep theorem by Jean Bourgain about how sets expand under multiplication in certain matrix groups.
Taken together, these two principles unify and extend a scattered collection of earlier results. They explain, through a single framework, why diverse combinatorial sequences all display similar spreading behavior among residues. The paper's contribution is both to organize existing knowledge and to push the bounds further for sequences like rows of Pascal's triangle and blocks of Catalan numbers, showing they achieve more distinct residues than a naive square-root estimate would predict.