I enjoyed reading Dan Fiorino's article but couldn't help feeling some of the same criticisms that he leveled against himself. Defense is equally as important as offense. Defensive players should be celebrated for their successes too. An all-encompassing stat should be very carefully tuned and weighted. Even then, there are intangables that people won't ever record on a stat-sheet. In the end the only thing that matters is that your team wins as many points as possible. Lucky for me Dan made the data publically available. This analysis aims to highlight a different subset of players - the players that quietly do all the little things right and grind for every point. Two caveats:
To capture how effective teams and players are on offense and defense I calculated:
No suprise that the Madison Radicals had stellar offense and defense on their championship run. It is suprising that the Toronto Rush missed championship weekend entirely. In the west, the Cascades, Spiders, and Flamethrowers could only be carried so far by their offensive talent. The NY Empire, DC Breeze, and Growlers relied on their defense to break for the win.
Select a team below to see the efficencies of the individual players on that team who played 10 or more points.
In general, the more points a player tallied on offense, the poorer their play. On defense however, the team stalwarts performed better than the journeymen. This goes against the common sense that you must play a lot of offensive points to get chemistry with teammates and have a smooth flow. Also on defense I would have guessed the players who play most would get tired and less effective.
Teams with large positive slopes for their offensive / defensive linear regressions would likely benefit from playing their stars even more frequently. Those with negative slopes might advance by giving less-known players additional touches.
It's easier to play good defense when you're emersed in a strong system led by a visionary coach and surrounded by experienced, bought-in teammates. On the fledgling teams in the AUDL, players must find their own way, and the stand-outs should be all the more celebrated. Compared to the team linear regression, the following players had positive error's of prediction.
Note: Sample size is key.
On offense, Kiran Thomas, Manuel Eckert, Nick Murphy, Joe Thompson, David Snoke, and Joey Huppert stood out with only 50-100 points played.
Players like Cam Bailey, Matt Auletta, Zach Sabin, Sam VanDusen, Dalton Smith, Sam Sohn, Keegan North, Matt Stevens managed to stay on the plus side while shouldering a major share of their teams possessions.
Less utilized defensive overachievers include Daniel Sperling, Pawel Janas, Terrance Mitchell, Michael Tran, Jack Williams, Max Sheppard, Bobby Ley, Kyle Stapleton, and Travis Dunn.
Patrick McMullen, Delrico Johnson, Brad Scott, Xavier Payne, Dylan Declerck, Iain Mackenzie, Shane Worthington, Nick Mahan, Mike Drost, Dillon Larberg, and Matt Lemar excelled throughout lots of defensive possessions.
Thanks for reading! I welcome feedback at afbcary@gmail.com. The code that generated this article is open source at https://github.com/afbcary/audl-stats. Help me find a better stat than error of prediction for the last two graphs (Cook's distance or leverage?).