| Abstract: | The aim of this paper is to gather insight into typical-case complexity of the Boolean Satisfiability (SAT) problem by mining the data from the SAT competitions. Specifically, the statistical properties of the SAT benchmarks and their impact on complexity are investigated, as well as connections between different metrics of complexity. While some of the investigated properties and relationships are "folklore" in the SAT community, this study aims at scientifically showing what is true from the folklore and what is not. |
