From a mathematical perspective, prediction markets are a zero-sum payment system that heavily relies on scoring rules and convex cost functions. However, most operators in reality only implement the addition and subtraction logic of linear funding pools, making it impossible for the market to effectively discover probabilities. Instead, it systematically degrades into a path-dependent fund allocation tool. In plain terms, if you don’t even know what a convex function model is, how are you supposed to build a prediction market? At best, you’re just setting up another casino.