Comparing soccer—as well as basketball and hockey players—has often been an inexact science. Some refer to their approach as the "eye" test while others use more analytical approaches. For instance, basketball has "wins produced per 48 minutes played", hockey has "goals versus threshold" and soccer has "goal probability added". But one of the problems one data scientist poses is it is difficult to compare athletes who play different positions, have different styles of play and may emphasize different skill sets.
At the 2017 Sloan Sports Analytics Conference, Shael Brown, an undergraduate student at Dalhousie University in Canada, proposes a new metric based upon the PageRank algorithm, the way Google's search engine ranks websites in its results. His presentation discusses how players' "relative ranks" when it comes to the flow of their game, can better help us compare the best in each sport.
