Prediction Market Profits: Accuracy vs. Strategy Explored.

This research addresses a long-standing puzzle in prediction markets: why informed forecasters often lose money while less accurate strategies can profit, contradicting classical theory. The study establishes a formal equivalence between predictive accuracy and profitability, specifically for central limit order book exchanges, which are prevalent today. It introduces a "proper" betting strategy that depends solely on a forecaster's prediction and the market price. The core finding is that this proper betting strategy consistently generates positive expected profit whenever a forecaster's prediction outperforms the market price, provided sufficient liquidity exists. Furthermore, this strategy is identified as essentially the only one offering such a robust profitability guarantee. The proof relies on a generalized decomposition of expected profit, extending classical automated market maker (AMM) guarantees and explaining how profit can occur without an inherent accuracy edge. Empirical validation across thousands of AI model forecasts demonstrated that proper betting was the sole strategy reliably converting accuracy into profit. A live deployment on Kalshi, a prediction market platform, yielded an impressive +80.33% return on investment with a Sharpe ratio of 3.35 over a month. The research also identifies distinct forecasting personas and how optimal proper strategies vary among them.