Following the deadline for submissions, the Pro Forum judges spent the early weeks of this year reviewing dozens of presentation proposals for the 2022 event, which will be taking place on Wednesday 23rd March.
The judges scored each proposal, which were anonymised, based on four key criteria: innovation, methodology, relevance and application. By the end of the process, six projects had been chosen for presentation to an invited audience of industry delegates attending at 30 Euston Square in London and via an exclusive livestream.
Now in its ninth year, the Forum remains a key date in the football analytics calendar, providing presenters with a platform for showcasing innovative approaches to key questions relating to performance analysis and recruitment. A number of previous event presenters have gone on to work for leading clubs, as highlighted here.
For the second successive year, the event has partnered with the French Football League (LFP), who will be making historical Ligue 1 Uber Eats tracking data available to help presenters attain deeper insights into performance.
The full line-up for the 2022 Forum, listed in no specific order, is as follows:
Gerald Lim, Ashley See, Zhi Yuan Chua – Pressing Times: Can Data Tell Us When and How to Navigate Out of a Counter Press?
This presentation applies event and tracking data to identify different types of in-game scenarios involving a counter press, with the ultimate objective of establishing which scenarios, from a risk perspective, are optimal for the team in possession to play through a press.
The project will look to qualify and quantify counter pressing strategies through the application of four metrics, then compute the successes of various decisions taken against the counter press based on key outcomes including ball progression, retention time and the threat conceded from losing possession.
Gerald, Ashley and Zhi Yuan all studied together at Imperial College London, graduating in Physics (MSci), Computing (BEng) and Chemical Engineering (MEng) respectively. Zhi Yuan is still based in London, working as a Data Analyst for Deloitte UK, whilst Ashley works as a Software Engineer at the Centre for Strategic Infocomm Technologies in Singapore. Gerald is now a Research Scientist at the Centre for Climate Research Singapore.
Caterina De Bacco – Measurable KPIs for Scoring Efficient Runs
This project was chosen in the first of the Forum’s two club-led submission categories, set by Tyler Heaps of AS Monaco, which applies tracking and event data to distinguish proactive runs from reactive runs, with the objective of identifying players whose runs are the most effective in both an attacking and defensive context.
From a defensive standpoint, the model will label runs based on whether they result in attacking players being marked or closed down, or interrupt a potential passing line, whilst attacking runs will be credited based on whether they generate a passing opportunity, attack vacant space or trigger movement from an opposition player which in-turn creates space for a teammate.
By being able to separate proactive running from reactive running, using a score-based KPI system, performance analysts will be able to better contextualise the quality of the runs made by every player in a squad, to be assessed alongside longstanding metrics such as total distance covered and high intensity sprints.
This is the second successive year that Caterina has been chosen to present at the Forum, having delivered a presentation in 2021 on evaluating the efficiency of each player during the pressing phase. Caterina holds a PhD in Statistical Physics and is currently an Independent Research Group Leader at Cyber Valley, Max Planck Institute for Intelligent Systems in Tübingen, Germany.
Arnav Prasad – A Reinforcement Learning Approach to Scout Allocation and Talent Discovery in Football
Arnav Prasad is the winner of this year’s Dr. Garry Gelade award, which recognises the outstanding proposal submission from an undergraduate.
Arnav is currently pursuing a Bachelor of Science in Mathematics with a Specialization in Computer Science at the University of Chicago. His Forum presentation will address one of the major challenges faced by recruitment departments with a limited number of scouts. Using both empirical and simulated player data, Arnav’s work will seek to quantify the advantage that teams can gain in scouting via state-of-the-art reinforcement learning algorithms.
Having developed a simulation environment for this problem, Arnav will demonstrate how clubs can build and test various algorithms to inform their own strategic decision-making. This environment will also allow for experimentation with various player sampling distributions and noise settings.
In parallel with his studies, Arnav is currently working with the Boston Red Sox in their baseball analytics group. Arnav previously interned for the Los Angeles Dodgers in baseball analytics and Citadel LLC, a leading quantitative hedge fund, in quantitative research & trading. He was selected as a 2021 Yoseloff Scholar by the Society of American Baseball Research (SABR), Finalist in the 2021 MIT Sloan Sports Analytics Research Paper Competition, and Winner of the 2020 ESPN Hackathon at the MIT Sloan Sports Analytics Conference.
Abhishek Amol Mishra & Soumyajit Bose – Measuring and Modelling Defensive Efficiency With Only Event Data
In the second of the Forum’s club-led submission categories, set by Luton Town’s Jay Socik, Abhishek and Soumyajit’s project evaluates ways of judging the defensive ability of centre-backs by going beyond standard numbers derived from event data, such as tackles and interceptions.
Their presentation will focus on the application of both existing metrics, which have been possession adjusted, and new advanced metrics which focus on opposition goal, shot and threat prevention to identify standout central defenders in the EFL Championship, with the overall objective of enhancing the recruitment profiling of players as part of a club’s technical scouting process.
Based in Texas, Abhishek is an undergraduate studying Computer Science at The University of Texas at Dallas, whilst Soumyajit is currently completing a PhD in physics at Cornell University.
Bakr Annour & Silvio Matano – Identifying Ways to Efficiently Break a High Press
Bakr and Silvio’s project aims to present an innovative way of applying event and tracking data to generate formation and player recommendations based on the pressing styles adopted by an upcoming opponent, helping performance analysts identify the most effective way that a specific pressing style can be bypassed.
The presentation will outline how sequences of interest are identified and how their success is defined, explain how their model is used to ‘place them’ in a common space, and how the backend of a hypothetical application would interact with the model.
Their recommendations will be underpinned by insights derived from recurring event sequences and player attributes which have been successful against specific pressing styles and shapes, which can be applied to inform a tactical approach pre-match and to influence in-game decision-making from the bench.
Based on London, Bakr is a Senior Operations Analyst currently working in the logistics sector. Silvio also lives in the capital, where he works as a Statistical Officer in the public sector.
Ashwin Phatak, Henrick Biermann & Franz-Georg Wieland – Expected Counter: Probabilistic Modelling of the Occurrence and Danger of Counterattacks in Soccer Before They Occur in Real Time
This project was selected in the Stats Perform-led category, which aims to predict the success of a possible counter attack and its potential danger to the defending team, measured by a new metric: Expected Counter.
The presentation will build on existing work showcased by Luke Bornn and Javier Fernandez for measuring space creation, by adding additional variables including the game state, the on-field position of players on both teams, the location of the ball and their respective velocities.
Ashwin and Henrick are both based at the German Sport University in Cologne, working as Research Assistants in the Institute of Training and Computer Science. Franz-Georg Wieland is a Doctoral student and research associate at the University of Freiburg.
Stats Perform would like to thank everyone who submitted a proposal as well as the panel of judges, and congratulate the six people who will be presenting at the 2022 Pro Forum.
