[Paper Review] Leicester's Tale: Another Perspective on the EPL 2015/16 Through Expected Goals (xG) Modelling

Sheikh Badar Ud Din Tahir, Leonardo Egidi+1 more|arXiv (Cornell University)|Feb 17, 2026

Sports Analytics and Performance0 citations

TL;DR

The paper builds an inference-based xG framework to simulate EPL 2015/16 season standings, quantifying ranking uncertainty and rare outcomes, and examining mid-season xG as an early-warning signal.

ABSTRACT

Probabilistic modeling is an effective tool for evaluating team performance and predicting outcomes in sports. However, an important question that hasn't been fully explored is whether these models can reliably reflect actual performance while assigning meaningful probabilities to rare results that differ greatly from expectations. In this study, we create an inference-based probabilistic framework built on expected goals (xG). This framework converts shot-level event data into season-level simulations of points, rankings, and outcome probabilities. Using the English Premier League 2015/16 season as a data, we demonstrate that the framework captures the overall structure of the league table. It correctly identifies the top-four contenders and relegation candidates while explaining a significant portion of the variance in final points and ranks. In a full-season evaluation, the model assigns a low probability to extreme outcomes, particularly Leicester City's historic title win, which stands out as a statistical anomaly. We then look at the ex ante inferential and early-diagnostic role of xG by only using mid-season information. With first-half data, we simulate the rest of the season and show that teams with stronger mid-season xG profiles tend to earn more points in the second half, even after considering their current league position. In this mid-season assessment, Leicester City ranks among the top teams by xG and is given a small but noteworthy chance of winning the league. This suggests that their ultimate success was unlikely but not entirely detached from their actual performance. Our analysis indicates that expected goals models work best as probabilistic baselines for analysis and early-warning diagnostics, rather than as certain predictors of rare season outcomes.

Motivation & Objective

• Quantify ranking uncertainty in a full EPL season using xG-driven simulations.
• Assess how well xG-based models reflect actual league structure and final standings.
• Identify and characterize rare or extreme season outcomes within a probabilistic framework.
• Evaluate the ex ante diagnostic value of mid-season xG information for predicting second-half performance.

Proposed method

• Develop a logistic regression-based xG model using hand-crafted shot-level features (location, distance zone, shot_place, bodypart, situation, assist_method, fast_break).
• Compare three model specifications (Base distance zones, Distance Zone × Body Part interaction, Granular distance zones) using AIC and residual deviance to select the best fit.
• Estimate xG for shots, then feed team-level xG into a Poisson-based process to simulate season outcomes and standings.
• Conduct 1000 season simulations to derive expected points, ranks, and probabilities of outcomes like winning the title or relegation.
• Display mid-season diagnostic analysis by using first-half xG to simulate second-half results and assess predictive value.

Figure 1 : End-to-end xG-based framework for season-level ranking simulation and inference

Experimental results

Research questions

• RQ1 Can an xG-driven probabilistic framework reproduce the EPL 2015/16 league structure and top/bottom distinctions?
• RQ2 What is the extent of ranking uncertainty for teams when comparing xG-based simulations to actual final standings?
• RQ3 How informative is mid-season xG for predicting second-half performance and rare outcomes like Leicester’s title?
• RQ4 Do xG simulations provide meaningful ex ante probabilities for winning the title, top-four finish, or relegation?
• RQ5 What are the limits of xG in forecasting rare season events and how can they aid early warning diagnostics?

Key findings

• The granular distance-zone model provides the best fit (lowest AIC and residual deviance) among the tested specifications.
• Mid-season xG provides a monotonic signal for second-half performance, indicating xG captures underlying performance beyond current points.
• Leicester City ranks among top teams by mid-season xG and has a measurable, non-negligible title probability, though their eventual title is a low-probability outcome under the mid-season information.
• Across 1000 simulations, the framework reproduces the broad league structure and generates plausible distributions of final outcomes, highlighting ranking uncertainty and the rarity of extreme results.
• The approach emphasizes probabilistic baselines and early-warning diagnostics over deterministic prediction of rare season outcomes.

Figure 10 : Convergence of average predicted ranks with increasing simulations