Using PageRank to rank sports teams
http://carlmeyer.com/pdfFiles/SASGF08RankingPaper.pdf
In the article “Generalizing Google’s PageRank to Rank National Football League Teams”, the authors discuss a new application of the PageRank algorithm for ranking football teams. In this algorithm, you create a graph of all NFL teams and point an arrow from team a to team b if a lost to b, and then applying PageRank. The algorithm also weights the edges based on the margin of victory of that game. Conventional methods of ranking football teams simply depends on how many wins and losses a team has, which often results in ties. It can be unfair to compare teams whose early games were against strong opponents to those who first played easier opponents. Thus, this algorithm is a more nuanced way to rank NFL teams throughout the football season, as the strength of the teams that a particular team won and lost to really matters.
This applies to this class, as we have been discussing the PageRank algorithm. I think that considering the NFL as a network is really interesting, and makes a lot of sense even if few people would normally think of it in this way. This algorithm could also be applied to any sports league, making this application more broad. As a casual fan of NFL football, I found this to be a fun application of class material, especially since I have found that the conventional method for ranking teams is often frustrating, particularly when my favorite team (the ravens) or another team I like is not doing so well early on because they went up against some very strong rivals.