Data-Driven Ghosting using Deep Imitation Learning


Hoang M. Le (California Institute of Technology)
Peter Carr (Disney Research Pittsburgh)
Yisong Yue (California Institute of Technology)
Patrick Lucey (STATS)

MIT Sloan Sports Analytics Conference 2017

March 3, 2017

Motivated by the original “ghosting” work, we showcase an automatic “data-driven ghosting” method using advanced machine learning methodologies applied to a season’s worth of tracking data from a recent professional league in soccer. An example of our approach is depicted in Figure 1 which illustrates a scoring chance that Fulham (red) created against Swansea (blue). Suppose we are interested in analyzing the defensive movements of Swansea. It might be useful to visualize what the team actually did compared to what a typical team in the league might have done. Using our approach, we are able to generate the defensive motion pattern of the “league average” team, which interestingly results in a similar expected goal value (69.1% for Swansea and 71.8% for the “league average” ghosts — to fully appreciate the insights revealed by data-driven ghosting, we urge the readers to view the supplemental video.

Download File "Data-Driven-Ghosting-using-Deep-Imitation-Learning-Paper"
[pdf, 2.16 MB]

Copyright Notice

The documents contained in these directories are included by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a non-commercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.