Login / Signup

Developing a two-dimensional landscape model of opportunities for penetrative passing in association football - Stage I.

Pedro J M PassosRodrigo Amaro E SilvaLuis Ignacio Gómez-JordanaKeith Davids
Published in: Journal of sports sciences (2020)
This study investigated a method for modelling a landscape of opportunities for penetrative passing completed on the ground by ball carriers in association football. Analysis of video footage of competitive, professional football performance was undertaken, identifying a sample (n = 20) of attacking sub-phases of gameplay which ended in a penetrative pass being made between defenders to a receiver. Players' relative co-positioning during performance was modelled using bi-dimensional x and y coordinates of each player recorded at 25 fps. Data on player movements during competitive interactions were captured using an automatic video tracking system, recording player co-locations emerging over time, as well as current and estimated running velocities. Results revealed that the half spaces between the midfield and both sidelines were the key locations on field providing most affordances for penetrating passes in the competitive performance sample analysed. Due to the dynamics of players' co-adaptive performance behaviours, it was expected that opportunities for penetrative passing by ball carriers would not display a homogeneous space-time spread across the entire field. Results agreed with these expectations, showing how a landscape of opportunities for penetrative passing might be specified by information emerging from continuous player interactions in competitive performance.
Keyphrases
  • high school
  • single cell
  • deep learning
  • machine learning
  • electronic health record
  • high intensity
  • big data
  • mass spectrometry