Welcome back to the next edition of our collaboration with Zack Capozzi over at Lacrosse Reference. Each week, we are putting together a stat of the week to highlight a game, team, or individual from Men’s or Women’s lacrosse to focus on.
It might be from a game the previous weekend, or it might be for an upcoming matchup. But either way, it will allow me to use some stats combined with a little bit of commentary for why I think that particular one jumps out to me in a given week.
This week’s focus: Strength of Record vs RPI
Since we’re now in late April, all eyes are looking forward to Selection Sunday in a few weeks that will determine which teams will have a chance to play for a National Championship and which will not. Because of this, a key metric starts floating around in the lacrosse Selection Sunday conversation: RPI. RPI stands for Rating Percentage Index, and it’s the standard for nearly every NCAA sport including lacrosse when determining seeding for tournaments.
It’s important to take a little bit of a detour down this rabbit hole for a second, so bear with me. The whole reason why college sports find themselves in this predicament of needing an invented metrics while most professional leagues do not is because there are just so many schools. In a professional league, you can rely on standings, divisions, etc. to determine where teams fall in the postseason. Some leagues also play so many games that everything sorts itself out pretty cleanly, and if they don’t it’s usually because schedules are relatively similar and teams almost all play each other.
But in NCAA sports you’re dealing with potentially hundreds of colleges and fewer than twenty games in most sports. That means there is a severe lack of overlap in schedules and many teams won’t even play each other, let alone have several common opponents. Enter RPI. RPI was developed to give selection committees a numerical way of ranking teams. The way it does that is by combining a team’s record with their opponents’ record, and then their opponents’ opponents’ records.
This method as you may imagine has created a fair number of critics. Strictly by looking at records, you still are not getting a true feel for how good a team may or may not be. It certainly isn’t perfect, but it is better than some alternatives out there. This lack of perfection also drove the NCAA to adopt NET rankings in basketball which takes much of the same information as RPI into account, but also adds margin, game location, efficiency, and win quality. At this time, this is still unique to basketball and just about every other sport relies on RPI.
So now let’s talk about the Lacrosse Reference Strength of Record (SOR). Stealing straight from their glossary.
“Strength-of-Record is calculated by comparing a team’s actual results to what a theoretical Top-10 team should achieve against the same schedule. A team with a very difficult schedule may not have a gaudy win-loss record, but neither would a theoretical top-10 team if the schedule is hard enough. Using SOR allows us to account for the strength of a team’s opponents and factor in whether they actually won or lost those games in a way that a simple Strength-of-Schedule metric doesn’t.”
What I thought would be interesting is taking a fully record-based approach and comparing it to the SOR values. On top of that, we also have the lacrosse ELO rankings to see which teams overall are ranked the highest. So what you have below here are those three metrics all listing their top 25 schools.
Rank | RPI | SOR | ELO |
1 | Maryland | Maryland | Maryland |
2 | Princeton | Georgetown | Georgetown |
3 | Georgetown | Princeton | Yale |
4 | Penn | Rutgers | Cornell |
5 | Yale | Cornell | Duke |
6 | Rutgers | Virginia | Rutgers |
7 | Virginia | Army | Virginia |
8 | Cornell | Jacksonville | Army |
9 | Brown | Yale | Princeton |
10 | Ohio State | Ohio State | Denver |
11 | North Carolina | Duke | Notre Dame |
12 | Duke | North Carolina | Saint Joseph’s |
13 | Denver | Boston U | Jacksonville |
14 | Boston U | Penn | Penn |
15 | Harvard | Saint Joseph’s | Richmond |
16 | Saint Joseph’s | Utah | Loyola |
17 | Notre Dame | Brown | North Carolina |
18 | Richmond | Notre Dame | Vermont |
19 | Utah | Harvard | Villanova |
20 | Army | Richmond | Brown |
21 | Johns Hopkins | Denver | Boston U |
22 | Jacksonville | Villanova | Ohio State |
23 | Villanova | Lehigh | Lehigh |
24 | Loyola | Bucknell | High Point |
25 | UMass | High Point | Drexel |
Now if we circle back to the original fact that RPI is what the NCAA Lacrosse Selection Sunday committee uses, we can look at this comparison table to see who may benefit from RPI compared to the other metrics, and who may be hurt by it.
Who benefits? Primarily it’s looking like the Ivy League, but specifically Penn and Princeton. Princeton is all the way up to #2 in RPI and #3 in SOR, but falls to #9 in ELO. The #9 jumps out to me because that’s the difference between hosting a winnable first round NCAA game and going on the road to a tough matchup with an even tougher quarterfinal. The other team is Penn who is a lacrosse RPI all-star . They are all the way up to #4 in RPI, but sit at #14 in SOR and ELO. That’s bordering on hosting a game as a high seed via RPI and barely making the tournament via the other metrics.
Who is hurt by RPI? Jacksonville is the team I see with the worst RPI, solid ELO, and great SOR. A #22 RPI would suggest they have to win the SoCon to make the tournament. But everything else in the world says this is a tournament team. The other one is Army who is just ahead of Jacksonville at #20 in RPI, but is a top eight team in SOR and ELO. That’s another example where all of a sudden RPI is saying that Army has to win the Patriot League AQ to guarantee a spot in the tournament.
What are your thoughts looking at these? They’re still a snapshot in time and these numbers change with every single game that’s played. But which one jumps out to you as “right”? Or are things still a little bit off from what you would expect to see?