The article reviews methods for estimating statistics without assuming a specific mathematical form. It focuses on practical applications in economics rather than just theory. The researchers discuss ways to estimate the shape of data distributions, predict outcomes without strict assumptions, and combine both flexible and fixed models.