Since the first staging of the Forum in 2014, dozens of data practitioners have earned the opportunity, through the event’s analytics competition, to present innovative new approaches for deriving performance insights using data to invited delegates from clubs and national federations.
Next year’s Opta Forum, taking place on Tuesday 21st March, will offer another generation of analysts the opportunity to showcase their concepts to the wider industry, with initial proposals being accepted until 20th December.
Ahead of this deadline, we have taken the opportunity to dig into the archives and revisit a number of previous Forum presentations, delivered by people who are now working across the football industry.
Abel Lorincz, 2015 – Breaking Down the Possession Game of FC Bayern Munich
At the second edition of the Forum, Hungarian analyst Abel Lorincz delivered a presentation exploring how graph theory and connectivity measures could be used for finding the weakest players in a team’s possession game, focusing specifically on the approach of Bayern Munich.
At the time Abel was six months into a five year spell working for Budapest Honvéd, initially as an academy video analyst before progressing to working as a first team analyst.
Today he works in the Swedish Allsvenskan, in the role of Assistant Coach of Analysis at Hammarby Fotboll.
Sam Gregory, 2016 – The Element of Surprise in Attack
A project chosen in 2016 was devised by Canadian analyst Sam Gregory, who was inspired by approaches in the NFL to deliver an in-depth analysis of the potential benefits of including a surprise attacking play as part of a team’s varied approach going forward.
After co-founding a football analytics consultancy, Sam spent a number of years working for different industry service providers, including Stats Perform, prior to taking up the role of Director of On-Field Analytics at York 9 in the Canadian Premier League. Prior to the start of 2021 MLS season he was appointed Director of Analytics at Inter Miami, where he supports front office, technical staff and coaching staff in recruitment, performance analysis and sport science.
William Spearman, 2017 – Football Passing: Physics Based Modelling of Pass Probabilities in Football
The work of William Spearman was chosen for presentation at the Forum for two successive years. After delivering a presentation on pitch control in 2016, William returned in 2017 with another project which applied event and tracking data to assess the likelihood of a pass being completed, allowing clubs to better evaluate passing skill.
A year after that presentation, William joined Liverpool FC as the club’s Lead Data Scientist, a role he continues to hold on Merseyside.
John-Mark Sisman, 2018 – An Analytical Approach to Exploring Pressing Styles
In 2018, bioanalyst John-Mark Sisman introduced a new approach, using tracking and event data, to identify a team’s approach to pressing. The project looked to identify pressing triggers, based on different scenarios when the opposition team had possession, taking into account various different contextual possession factors.
In 2020 John-Mark joined the City Football Group in the role of Performance Physicist. Based on his experiences of presenting at the Forum, he has the following advice for anyone who is interested in sending in a submission for next year:
Joe Mulberry, 2019 – Gazing into Latent Space
After first presenting at the Forum in 2017, Right to Dream and FC Nordsjælland’s Joe Mulberry returned two years later to deliver a presentation focusing on the categorisation of possession sequences.
His presentation highlighted how analysing sequences, using a combination of an auto-encoder and T-SNE, can allow their categorisation using their location within latent space.
Joe was one of two presenters in 2019 who were already working in full-time positions at clubs when their work was selected for presentation, demonstrating how the event was by now attracting a growing number of submissions from industry practitioners alongside the wider analytics community.
Karun Singh, 2020 – Learning to Watch Football: Self-Supervised Representations for Tracking Data
Karun Singh presented at the Forum in 2020, outlining a novel approach to identifying recurring match scenarios using tracking data.
Through applying deep learning processes to the tracking dataset, Karun’s presentation explained how by leveraging ‘self-supervised’ labels, AI technology can help match analysts quickly identify similar scenarios, such as counter attacks, across multiple games without having to scour through hours of video, enhancing their working processes.
At the time of his presentation Karun was working for Facebook as a Software Engineer, before moving onto Grammarly where he spent nearly two years working as a Machine Learning Engineer. In October 2022 he was recruited by Arsenal, where he works for the Premier League club as a Data Scientist.
Vignesh Jayanth, 2021 – Identifying and evaluating strategies for successfully penetrating a high opposition press from short goal kicks
At the 2021 virtual event, Vignesh Jayanth was chosen to present for a second successive year in one of the Forum’s club-led submission categories.
Responding to a question put forward by Lucy Rushton, Vignesh’s presentation focused on an approach to identify and evaluate the best-attacking strategies to penetrate a high opposition press from short goal kicks.
Five months after presenting at the Forum he was recruited by Ligue 1 side Stade Rennais as a data scientist. He will be involved with the Opta Forum again in 2023, putting forward his own reseach question in one of the practitioner-led submission categories. Vignesh will also mentor the selected candiatate.
Soumyajit Bose, 2022 – Measuring and Modelling Defensive Efficiency With Only Event Data
Together with Abhishek Amol Mishra, Soumyajit Bose presented at the most recent edition of the Forum, presenting the findings of a project chosen in a club-led submission category set by Luton Town’s Jay Socik.
Abhishek and Soumyajit’s project evaluated 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 focused on the application of both existing metrics, which were possession adjusted, and new advanced metrics which focused on opposition goal, shot and threat prevention to identify standout central defenders in the EFL Championship.
In the months following his presentation, Soumyajit followed in the footsteps of Sam Gregory and was recruited to Stats Perform’s AI team, working as an AI scientist.
If you are interested in following in the footsteps of these presenters and would like to submit a proposal for the 2023 Opta Forum, check out the various submission categories and criteria. Full details can be found here.
For more information on the 2023 Opta Forum, please contact email@example.com.