Form is temporary, class is permanent: Meeting report

The event ‘Form is temporary, class is permanent: Statistical methods for predicting the career trajectories and contributions of Test cricketers’ was a webinar held by the Statistics in Sports section of the Royal Statistical Society on Wednesday 16th June, 2021 at 10am (BST) via Zoom. The speaker was Oliver Stevenson, currently a Data Scientist at Luma Analytics in Auckland, New Zealand. Having recently completed his PhD at the University of Auckland investigating new statistical methods and models in the field of sports analytics, Oliver presented an informative and entertaining talk about modelling the performance of batsmen to include stable characteristics as well as the more common recent form. The event was organised by Pete Philipson, one of the committee members and garnered an online audience of 42.
The presentation was accessible to those who know little about cricket but are interested in modelling as well as to cricket enthusiasts with less well-developed statistical skills. Oliver’s friendly and conversational style captured and held the attention and provided clarity about the decisions behind the models. The mathematics were presented, although the flow of the talk did not depend on a full understanding of the modelling process. Throughout, the methods were applied to particular players to demonstrate how form fluctuates against an underlying continuum. 
The discussion after the talk was lively, with the subject matter provoking particular interest amongst the cricket fans in the audience. The talk was held immediately after a Test series involving England and New Zealand, and a week prior to the inaugural World Test Championship, for which the speaker made a prediction based on his work. This added relevance and currency to the event.
Report prepared by Dr Pete Philipson, Senior Lecturer in Statistics, Newcastle University.
Pete’s current research encompasses computational methods for faster fitting of multivariate joint models of longitudinal and event time data as well as software development (implemented in R) for multivariate joint modelling of longitudinal and time-to-event data.
Load more