Using PageRank to Determine Tiger Woods’ All Time Golf Ranking
One of the biggest debates in every sport is determining the best player to ever participate. However, over the course of the past decades, the gameplay, competitors, and many other factors change the way that a player is successful in his or her respective sport. In this article written by Daniel Brannock, a data science consultant for Elder Research, he attempts to compare Tiger Woods with other professional golf players who are considered to be the best to ever play. To begin the analysis, Brannock uses graphs connecting players based on their placing in professional golf tournaments that are considered “Majors”. Majors are the four main tournaments in golf, and are considered to be the most important to win out of all other tournaments. He uses data on 4000 golfers from 1970 to the present. By comparing all these players, he is able to see not only how all players faired in their tournaments, how long they remained in their prime, etc., but he is also able to measure the level of the competition in each time period. To get final results, Brannock cuts out about 80% of the initial golfers who haven’t performed as well as the golfing greats over the past few decades. He then uses the PageRank method to get a final score for each player.
To compare players and their results from the Majors, each tournament is measured the same way: the first place winner is connected in a network as a node to all other players in the tournament, who each have their own node. The weights on the edges between each player and the winner is the difference in their score. The 4000 golfers generate a little under 2 million edges. If PageRank is applied here, the best players of all time are not necessarily ranking the highest, so Brannon narrows the results to remove outliers. The qualifications for a player to make the final dataset was a career lasting twelve years or longer, and the only data about them is from their eight highest performing years. This ensures that players who had one great season and didn’t play much longer than that are not considered the best of all time. Furthermore, in golf, players can continue to play at a high level for decades, even though they won’t play at the same level as they did in their prime. By only including their eight best years, it doesn’t matter if a golfer played until he was 60 but didn’t win anything for decades leading up to retirement.
Since data on players from the 70s might exclude the beginning of their careers, and data on players from now will exclude data from the future, Brannock normalizes the data. Finally, Brannock applies the PageRank method to the remaining 687 golfers, to get the following results:
In his final conclusion, Brannock states that Tiger Woods is the best golfer of all time. Although his analysis may have faults (some younger competitors who haven’t played out their entire careers are rated pretty high up, for example, Rory McIlroy, ranked 6th and aged 30), his initial question was to determine if Tiger Woods or Jack Nicklaus is the best of all time, and his results have them at #1 and #2, as they are commonly known.
Article: https://www.predictiveanalyticsworld.com/patimes/asking-the-right-analytics-questions-and-whether-tiger-woods-is-better-than-jack-nicklaus/10539/