As we have reached the midway point of the 2024-25 D-III hockey season, it’s time to take a first glance at what’s to come at the end of the road: tournament time.  This year, we have numerous changes happening across the Division III landscape.  These have a direct effect on selecting which schools will participate in the postseason.  I thought that the Christmas break would be a great time to give a little context to the selection process as a whole, what came before, and how we can analyze this information going forward.

What is the NPI?

The NCAA Power Index (NPI) is a new ranking system instituted this year for all Division III sports by the NCAA Championships Committee.  It was created by stat-guru Tim Danehy, the creator of the now-defunct Pairwise system (more on that later).  The NPI is… complicated.  The nuts and bolts of it all are located in this FAQ the NCAA released when it became official.  My task is to make it a bit easier to understand for us on the hockey side.  Essentially, the NPI is supposed to be an improvement on the RPI (ratings percentage index).  The RPI was established in 1981 for D-I college basketball, and then spread to other divisions and sports, including D-III hockey.  This has been replaced by the NPI, which is supposed to “ensure that selection criteria is applied consistently to teams during selections,” via the NCAA.  In other words, this is to improve upon the math and keep things as objective as possible.  Let’s get into how each are calculated and how they differ.

RPI Rundown

RPI (Old System):

-25%: team’s winning %

-25%: team’s opponents’ winning %

-50%: team’s opponents’ opponents’ winning %

Basically, the RPI was a way to not only take your team’s record into account, but the strength of schedule that team played as well.  Granted, these are all determined by wins and losses themselves, not by scoring margin or game control. It’s strictly a win % formula.  Each game on a team’s schedule got a RPI “game score.” However, there were additional weights that were applied to these “game scores,” which we will talk about shortly.

The Arcadia women check in at #8 in the inaugural NPI ratings. (Photo by Hunter Martin Photography)

NPI Rundown

NPI (New system for D-III men’s and women’s ice hockey):

-25%: team’s winning %

-75%: team’s strength of schedule %

The NPI calculation seems more straight forward, but how do we calculate one’s SOS?  Is it just basically RPI with a fancy new name?  To answer that, a team’s strength of schedule is basically just the average NPI rating of all of their games played.  So, it is a self-referencing calculation.  But similar to the RPI, there are specific indicators that we have to take into account.

Each game a team plays gets in own NPI “game score” in a way.  Now, there are so-called “bad wins” and “good losses” that inherently will inflate or deflate a team’s rating.  These are taken out of the equation.  Let’s say you have a 22-5 record, but 6 of your wins are against bottom-feeder teams.  These wins would inherently lower your own resume, so they are taken out of the math altogether, giving you a better win %.

Minimum Wins

The “negative wins” concept shows up again in this specific criteria.  The reality is, in weaker conferences, there are schools that have several of these games on their schedule.  To help judge these teams more evenly with teams that schedule more challenging games, you must have at least 9 wins accounted in your NPI.  No matter how many “negative wins” you have, you need to have a minimum of 9 wins factored in.  For instance, if a 26-3 team has 21 “negative wins,” we have to count their top 9 wins in the calculation, even if it weakens their NPI number.  This is why it’s paramount for schools to schedule tough non-conference games, even if they are in a lesser league.

Home/Away Weights

Ice hockey has specific rules to better account for the sport’s quirks and features.  Home wins are only worth a 90% weight when calculating that specific game score.  Whereas an away win is worth 110% for a win.  On the other hand, losing on the road, has 90% of that particular game score, and a home loss is more costly, at 110%.  There are no weights applied to neutral-site games.  This is obviously acknowledging home ice plays a key role. (This was also part of the RPI, but with different weights).

Overtime Weights

Similar to above, any game decided in overtime has its own specific weights.  An OT win is technically 67% of a win, while an OT loss counts for 33% of a win.  Although shootouts are now mandated by the NCAA, all games that head to a shootout officially go in the books as a tie.  One important caveat, is that these OT weights go away come conference playoff time.  The rationale being, if it is 5v5 hockey rather than 3v3, it should be treated as a normally-decided game (This was also part of the RPI).

Quality Win Bonus (QWB)

A team’s individual game score can get a boost for a so-called “quality win.”  If you defeat a team that is high enough in the NPI rankings, you get a 0.5 multiplier to your game score.  For D-III hockey, if you have an NPI rank of 53 (top 25 for women’s hockey, top 34 for men’s hockey), you have that multiplier added to your victory.  However, these are not all created equal.  Meaning, a win against #6 in the country, will be a larger bonus applied to that game, than a win against #20, all things considered.  I have the math below to visualize it a little better.

-Win over a team with an NPI of 58, (#20 on the women’s side)

58-53 = 5 * 0.5 (QWB multiplier) = 2.5 is your Quality Win Bonus

-Win over a team with an NPI of 72, (#6 on the women’s side)

72-53 = 21 * 0.5 (QWB multiplier) = 10.5 is your Quality Win Bonus

Tim Coghlin’s Green Knights of St. Norbert are ranked #3 in the first edition of the Men’s NPI rankings, (Photo by Patrick Ferron)

Differences from Last Year’s Pairwise

With all this kept in mind, you can now officially access the NPI rankings as of this moment on USCHO, which has been verified with the NCAA.  There is no question there is a lot of hockey left to be played, and the data will be ever-changing until Selection Sunday.  As of this year, the Pairwise era is officially dead (2018-2024).  Many folks will be saying “good riddance” to the pairwise, but in my opinion, we should be reserving judgment until we officially see how the NPI evolves the rest of this season.

The Pairwise rankings “paired” every team against one another in three categories:

  • RPI ranking (outlined above)
  • Head-to-head record
  • Record vs. common opponents

Whichever team took at least two of these three won the comparison, and earned a PW point.  The teams were then listed in order of most PW points won via comparison.  This system caused a lot of outrage, with math being the sole reason teams are selected.  Prior to the 2018-19 season, a selection committee was used to choose teams based off of criteria.  Starting this year, there is no extra emphasis placed on head-to-head or common opponents, and it is all strictly based on the formula (albeit a new one).  So the NPI rankings are the clear cut “seed list” you may have heard of in March Madness talks.

Issues

Unfortunately, in our sport, there is a serious scheduling conundrum.  With an extreme lack of cross-over play between the East and West region, many have found that the RPI (or any math equation) should not be in charge of setting the postseason field.  I’ve looked at advanced calculations on how the math works, and the RPI does get more accurate when there is ample non-conference play.  Having seven conferences out east playing amongst themselves allows for a pretty fair ranking (most of the time.)  But when there are three conferences in the other region that have very little opportunity but to beat each other up, it’s nearly impossible to rank them along with the east teams when there is so little cross-region play.  It’s a hard task, but in my opinion, only using math does not provide a fair shot for all, with many west teams from extremely competitive conferences getting the short end of the stick in these calculations (rankings).

UW-River Falls and Gustavus, who have played one another in the last 4 NCAA tournaments, are sitting at #11 & #15 in the initial NPI women’s rankings (Photo by Jordan Modjeski)

Looking Ahead

As we head into the home stretch of the season, I hope this post helped clear up any potential questions on how the new 14 (men) and 12 (women) team fields will be selected and seeded this March.  I am looking forward to writing more articles in the next few weeks/months, where I will project the bid selections and what potential brackets could look like!  Also, I provided some resources that the NCAA released regarding the NPI and NCAA tournaments at general.  If you’re curious, they provide some great information.  As always, please do not hesitate to reach out, as I’m more than happy to take any questions or comments down below!

← Back

Thank you for your response. ✨