Researchers have developed two novel methods for converting betting odds into accurate outcome probabilities, crucial for sports forecasting and market efficiency analysis. The first, an odds-only approach called OO-EPC, adjusts for bookmaker biases by assuming equal profitability confidence across outcomes, outperforming existing methods on a large football match dataset. The second, a generalized linear model named FL-GLM, incorporates historical data and specifically accounts for the favorite-longshot bias, offering a more interpretable and effective alternative to current models. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
RANK_REASON The submission is an academic paper on arXiv detailing new statistical models for sports outcome forecasting.
