Leveraging Minute-by-Minute Soccer Match Event Data to Adjust Team’s Offensive Production for Game Context
*, Ahmet Cemek, David Gillman · 2025
Read Leveraging Minute-by-Minute Soccer Match Event Data to Adjust Team’s Offensive Production for Game Context by *, Ahmet Cemek, David Gillman (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-04008.txt.
Introduction
Andrey Skripnikov, Ahmet Cemek, and David Gillman adjust soccer attack statistics for game context in this 2025 paper. Raw shot counts can mislead. A team that trails late will attack more, while the leading team sits back. England against France at the 2022 World Cup is given as an example.
The team used minute by minute event data from 15 seasons in five major European leagues. Their model accounts for score, red cards, home status, pre match odds, and match minute. Adjusted output is projected onto one standard setting, a tied home game with equal numbers. That lets readers compare attack strength fairly across teams and matches, without the distortion of game state.
Catalog record
| CREATOR | *, Ahmet Cemek, David Gillman |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 74 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-04008
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Leveraging Minute-by-Minute Soccer Match Event Data to Adjust Team’s Offensive Production for Game Context *, Ahmet Cemek, David Gillman Abstract In soccer, game context can result in skewing offensive statistics in ways that might misrepresent how well a team has played. For instance, in England’s 1-2 loss to France in the 2022 FIFA World Cup quarterfinal, England attempted considerably more shots (16 to France’s 8) and more corners (5 to 2), potentially suggesting they played better despite the loss. However, these statistics were largely accumulated when France was ahead and more willing to concede offensive initiative to England. To explore how game context influences offensive performance, we analyze minute-by-minute event-sequenced match data from 15 seasons across five major European leagues. Using count-response Generalized Additive Modeling, we consider features such as score and red card differential, home/away status, pre-match win probabilities, and game minute. Moreover, we leverage interaction terms to test several intuitive hypotheses about how these features might cooperate in explaining offensive production. The selected model is then applied to project offensive statistics onto a standardized ”common denominator” scenario: a tied home game with even men on both sides. The adjusted numbers - in contrast to regular game totals that disregard game context - offer a more contextualized comparison, reducing the likelihood of misrepresenting the relative quality of play. [1]Andrey Skripnikov 1 Introduction In soccer, commonly referred to as football, statistics play an important role in analyzing various aspects of the game—from assessing individual player and team performances to providing valuable strategic insights. For instance, metrics such as shot attempts and corner kicks can indicate how offensively dominant a player or team was during a match. However, relying solely on aggregate statistics without considering the broader context of game events may lead to misleading conclusions. Consider the 2022 FIFA World Cup quarterfinal between England and France (where ”FIFA” stands for Fédération Internationale de Football Association). England recorded 16 shot attempts compared to France’s 8, and earned 5 corner kicks to France’s 2, yet still lost 1–2 [ 1 ] . Based on these numbers alone, one might conclude that England played a more dominant offensive match despite the loss. However, these raw totals ignore the context of the scoreline: during the 40 minutes when the match was tied, France led in the aforementioned categories—7 to 5 in shot attempts and 2 to 0 in corner kicks. In the remaining 66 minutes, with France in the lead, they purposefully ceded possession and initiative to England as a tactical decision, focusing their efforts on defense and protecting the lead. The goal of this study is to develop a statistical adjustment that accounts for crucial contextual factors—such as score differential—when interpreting offensive metrics like shot attempts and corner kicks. In particular, we explore how various contextual variables affect offensive production, and subsequently project those effects onto a standardized ”common denominator” scenario: a home game where the score is tied and both teams are at even strength throughout. By doing so, we aim for the statistical outputs to more accurately reflect teams’ relative offensive performances, accounting for the likelihood of specific tactical adjustments under different game scenarios and reducing the risk of misleading conclusions. Additionally, leveraging minute-by-minute data allows us to test several hypotheses regarding how the impacts of various contextual factors may vary depending on the game minute—particularly when comparing early-game versus late-game situations. In this section, we will first review existing metrics in soccer analytics, along with related research in other sports. We will then outline our research question and emphasize its contributions to the field. 1.1 Background and Literature Review The field of soccer analytics has been growing over the recent years, including the strong increase in scientific publications associated with it over the past decade [ 6 ] . Various metrics and methods are well established in soccer performance analysis, designed to capture the complexities of game dynamics. One of the most widely adopted models in modern soccer analytics is Expected Goals, commonly referred to as xG [ 9 , 5 ] . This model estimates the probability that a given shot will result in a goal by considering factors such as shot distance, location, angle, whether it is a header, and the number of defenders between the attacker and the goalkeeper, among others. By quantifying the quality of scoring chances with numerical values, xG provides a method for assessing how effectively teams create goal-scoring opportunities. Beyond shot characteristics, some models also incorporate variables that serve as proxies for psychological influences, such as match attendance, game importance (based on tournament stakes), and goal differential [ 17 ] . With regard to similar metrics in other sports, Macdonald 2012 introduced a model to calculate expected goals in hockey based on statistics such as faceoffs, hits, and shots, facilitating performance evaluation at both the team and player levels (using adjusted plus-minus for expected goals). In rugby, Kempton et al. 2016 derived an expected point value for a possession by considering contextual factors such as field location and the outcome of the preceding possession. In American football, Yurko et al. 2019 developed a model to calculate added expected points—and even added win probability —for each play, accounting for game context such as down and distance, field position, and remaining game time, among other factors. Although extensive research exists on Expected Goals and analogous models—such as Expected Pass ( xPass ) [ 21 ] , Expected Threat ( xT ) [ 24 ] , and other possession value models [ 22 ] —the soccer analytics literature remains relatively sparse when it comes to statistical adjustments of existing metrics for game context. The closest example to this work is , where offensive statistics accumulated over time intervals during specific game contexts were similarly projected onto a baseline scenario. While the core underlying idea is the very similar, our approach employs a far more granular, minute-by-minute game event dataset, allowing us to capture potential shifts in game dynamics between earlier and later stages of a match—something Cemek et al. 2025 does not address, as it omits the timing of events. Our modeling framework is also considerably more robust and flexible, leveraging Generalized Additive Models instead of basic dummy-variable encoding for scoring and red card contexts. Among other works involving statistical adjustments in soccer, a notable example is the use of bivariate Weibull distributions to model the relationship between goals scored and conceded, as seen in Boshnakov et al. 2017 . Although these models are applied to soccer, they still often draw from frameworks developed in other sports—such as the ”offense-defense” model introduced by Harville 1977 in American football, where a team’s scoring output is modeled based on both its offensive ability and the opponent’s defensive strength. Additionally, acknowledging the impact of home-field advantage [ 8 ] , these models typically include adjustments for whether a team is playing at home. 1.2 Current Research Question This paper aims to address the limitations of traditional soccer statistics such as total shot attempts and corner kicks. Although not as pivotal as goals or goal-related metrics—the primary focus of most soccer analytics research (e.g., xG , xPass , xThreat )—these complementary statistics can still help construct a narrative about which