I built a soccer Elo model, put it in your browser, and then spent most of my effort trying to prove it could beat the bookmaker’s closing line. It can’t — and watching it fail taught me more about betting markets than any win would have.
Twenty-five Pine strategy files, dozens of indicators, and almost none of them were wrong. They were real patterns that didn’t survive a different window, a realistic cost, or an honest question about where the money came from. A catalog of how each one fooled me — sorted by failure mode, not by indicator.
My backtest showed +121.82% in three months, beating buy-and-hold. I spent real money — and a purpose-built validation pipeline — finding out whether that was real. It wasn’t. Proving it was the best engineering work in the project.