Give Me The Rock
this is fantasy basketball
Have you ever wondered what a visual representation of NBA players would look like using a two-dimensional scatter plot? Probably not (unless you’re a strange dude like me), but that’s exactly the question I asked myself the other day.
Using a statistical technique called multidimensional scaling (MDS), I collapsed the nine traditional statistical categories used in fantasy leagues (PTS, REB, AST, STL, BLK, 3PT, TO, FG%, FT%) for each player into two dimensions, which could then be plotted on a chart. MDS does the heavily lifting statistically, but it doesn’t have a clue conceptually what the dimensions are. I took a shot at naming them based on what I saw: a move from big men to small men going from left to right and the way all-stars rose to the top of the chart while role players fell to the bottom.
How does MDS work? The further two players are away from each other on the chart, the more dissimilar they were this season. According to the results below, the two most dissimilar players in the league this season are probably Dwight Howard and Kevin Martin. Conceptually that makes sense. In addition to having completely different games, Howard is the league’s worst free throw shooter, Martin’s is the league’s best.
Because the chart is small, I broke it out by position to help with readability. You can adjust the position using the slider below the chart. You can also view the chart by team. There is a ‘show history’ button that will show the cumulative results as you move along positions. It can be unclicked if the chart gets too cluttered. Finally, you can mouseover any dot on the chart to get the player associated with that item.
Just looking at the centers, you can see that Dwight Howard deserves to be in his own separate chart, preferably on another planet away from everyone else. Other than Howard, Andrew Bogut is the most “centerish” center, while Andrea Bargnani comes off as the most guard-like. I think that means this chart passes the smell test.
A note that stats are from the 2010-11 season and are based on per game averages. Players who didn’t finish in the top 200 for the season overall were also removed, as were guys like Yao Ming who missed more than half their team’s games.
Tags: Dwight Howard, Kevin Martin, statistical
Subscribe via RSS
Pingback: Tableau Public « Patrick Madden