2009 E-Lacrosse Feature: New lacrosse rankings systemOn Feb. 25, E-Lacrosse, the publication for which I am the editor and Web master, debuted the RRRR or Responsible, Reasonable, Reliable Rankings of Division I men's college lacrosse. This is a ranking but not a poll. I'll explain why it's different but generally the method is more inclusive, fair and accurate than typical polling schemes. Each Monday I will publish the RRRR Top 25 here on the blog with exclusive notes on that week's process. I'll also include my opinions about the teams on the move and the week's relevant matchups. We'll link to the full RRRR that "ranks" all 57 Div. I men's teams and fully explains the methods used.
An easy way to understand the RRRR is to equate it to the grading curve on a test. The "breaks" in this curve are where we feel a distinct and discernible difference in quality between two teams or two groups of teams. We create those breaks where they belong, based on the data, not in set groups. This week has a 1-3 "range" that contains Maryland, Hopkins and Duke. This weekend Maryland plays Duke and Hopkins plays a competitive game with Princeton, so we can expect the top group to split or at least shrink as the Maryland-Duke loser drops and the Hopkins-Princeton result is digested. Big wins by teams in the next range could move them up as well, changing the "volume" of each range.
Within each range, each week, teams are ordered by estimated quality and while no number is assigned, an order can be viewed. These are just the leanings of the participants in our process but the differences between any teams in the same range are not statistically significant enough to be more than that. This is where opinion, which in mathematics is often called "noise," is separated from opinion data, which is referred to as "signal". A signal is worth listening to. We've eliminated the noise but left the trace of it that fans find interesting. It's like showing our work. We'll show you even more here on the blog and I think you will find it very interesting.
This system is more fluid than other polls and allows room for uncertainty. More importantly it relieves the participants from being bound by earlier positions they have taken in the polling. To start with, we have no preseason "picks" that seem to be tacked (a finance term) to the results of most polls all year long, in every sport. Unlike most polls, the RRRR does not rely solely upon coaches or media members who have not seen many of the teams in person or on television, a common scenario. It's not that participants in other polls are lying to help teams they like or too lazy to investigate each of 57 Div. I teams. But they are forced to rely on guess work on most of their picks each week and it's not fair to them or the teams.
I would love to hear all of your comments on this new system. It is quite experimental. We invented it and are biased like the guy who picked U.Va. to be third to start this season. We hope you like it.
E-LACROSSE'S RESPONSIBLE, RELIABLE & REASONABLE RANKINGS (RRRR)
February 25, 2008
9-14 Notre Dame
February 22, 2008