← all strategiesWorld Cup Elo/Dixon-Coles Closing Line Value (CLV) Advisor
Statisticalscore 6found Jun 27, 2026github.com/KeaneYan/worldcup-polymarket-advisor
Generate fair-value World Cup match probabilities using Elo ratings and Dixon-Coles model adjusted for team form, then measure Closing Line Value to validate and improve model calibration.
How It Works
This read-only advisor (no live trading) computes fair-value win/draw/loss probabilities for World Cup matches using two established sports modeling frameworks: Elo ratings (a dynamic strength ranking system updated after each match result) and the Dixon-Coles model (a Poisson-based goals model that corrects for low-scoring game bias). These probabilities are compared against Polymarket market prices to identify potential edges.
Closing Line Value (CLV) Tracking
CLV measures whether bets made early in a market's life were at better prices than where the market ultimately settled before resolution. Consistently beating the closing line is considered strong evidence of genuine edge in sports betting research. This advisor tracks CLV as a model validation metric, allowing users to assess whether their model outperforms the crowd's final consensus.
Team Form Adjustment
Beyond static Elo, the model incorporates recent team form — a dynamic adjustment that weights recent match results more heavily. This captures momentum and injury/squad effects that historical Elo ratings may lag on.
Implementation Notes
Paper trading mode and scheduled automated reports are included. The tiny repo size (58KB) suggests a clean, focused Python implementation without heavy dependencies. Read-only design means it functions as a decision-support tool rather than an autonomous executor.
Risks
World Cup markets are heavily traded and likely efficient. The Elo and Dixon-Coles models are publicly known — any edge they offer may already be priced in by sophisticated participants. Tournament-specific factors (referee decisions, VAR, weather, squad rotation) are difficult to model. CLV tracking is only meaningful over large sample sizes, which a single World Cup cannot provide.