On 2 March, the RSS Merseyside Local Group hosted an event featuring talks on the topic of 'Statistics and football' to coincide with the recommencement of spectator sports, including the delayed Euro 2020. This was the fourth event to be broadcasted live via the RSS Merseyside YouTube channel and is now available to watch.
Questions were taken for the speakers from viewers watching the broadcast live using the chat feature. The event received 80 registrations and the video has received over 100 views total so far.
The first speaker, Professor David Firth (University of Warwick) delivered a talk on 'Schedule-adjusted league tables during the football season'. Professor Firth discussed the problems with standard league tables, some of which have become more evident through interruptions related to COVID-19. He then presented alt-3, a league tabling which incorporates strength of opponents played and home advantage. The mathematical basis of alt-3 (an advancement on Bradley-Terry models) was explained. Professor Firth finally demonstrated how informative this tabling is through data visualisations of results from Liverpool and Everton football clubs.
The second speaker, Dr Rob Mastrodomenico (Global Sports Statistics), asked 'Can we predict the future?' in 'An introduction to football modelling'. Dr Mastrodomenico’s talk gradually built up to a predictive model for match results, considering goals to follow a Poisson process that also captured home advantage and team ability. Applied to previous English football league data, this model showed that Manchester City’s 2014 Premier League win was quite predictable despite bookmakers’ speculation to the contrary! Dr Mastrodomenico used this to highlight the difference between market perception and statistical evidence.
These talks exemplified the diverse applications of mathematics and statistics to sport analytics, and an extended Q&A session followed, discussing the directions various sectors are taking in football modelling.
The RSS Merseyside Local Group plans to host two more meetings to take place in the last quarter of the year, subject to restrictions on in-person events.
Liam Brierley, MRC Skills Development Fellow, Department of Health Data Science, University of Liverpool