Bracket ABCs: Early Edition
by Adam Wodon/Managing Editor
There are number of sites out there that publish the Pairwise — the system that the NCAA uses to select the teams for the NCAA tournament. Until this year, they were all the same. Even though the committee doesn't publish it itself, the methodology for doing it was reverse engineered a long time ago, is fully explained in our Primer, and — other than the years the vague "bonus points" were in existence — everyone therefore had the same thing.
This year, the committee threw in some new factors that make things a little tricky — the weighting of home/road records, and a "Quality Win Bonus." And there isn't a history yet to fall back on.
Rest assured that the formula is still cut and dried, from the committee's perspective. There will be no guesswork or subjectivity involved. The committee's people will give them a final calculation at the end of the year based on the "right way," as defined by them, and that will be that.
But while the committee was transparent in how the weightings and Bonus were supposed to be done in general, it didn't completely explain how the numbers were supposed to be applied against the existing RPI. There are different ways to do it.
Therefore, different sites are showing slightly different results. And we've been fielding constant questions as to why ours doesn't match what's being shown elsewhere.
It gets pretty technical, but essentially, you can apply the weighting against the team's winning percentage only, or against the entire RPI (which consists of winning percentage, opponent's winning percentage, and opponent's opponents winning percentage).
There are also different ways of counting ties; the order in which things are calculated can make a difference; and, most importantly, which games are officially neutral site and which aren't is now of the utmost important. And in some cases, that's a question. For example, the Penn State-Robert Morris meeting in Pittsburgh was labeled "neutral," but RMU was designated the home team when it played Bowling Green.
All that said, we're confident ours is calculating it correctly. But we'll find out.
* * *
With the "Record vs. TUC" concept essentially folded into the RPI via the "Quality Win Bonus," the Pairwise chart tracks straight with RPI more than ever. There are only two other components now — Head-to-Head Record, and Record vs. Common Opponents. Well, if you don't play the team you're comparing to, that leaves only Common Opponents, but even if you win that, RPI is the tiebreaker, rendering the Common Opponents record meaningless.
Consequently, you have just two ways to flip comparisons — have a significant Quality Win Bonus (QWB) for beating good teams, or have a significant head-to-head superiority with the team you're comparing against. For example, Notre Dame is 15th in RPI and New Hampshire is 16th, yet they flip in the Pairwise chart because of UNH's recent sweep of the Irish.
As a result, projecting possible scenarios going forward is not as complicated as years past. Who you defeat is not as important as where — it's better to win on the road. There is no more "TUC Cliff" that causes teams' "Record vs. TUC" to change drastically from one day to the next, because the "Record vs. TUC" component was eliminated. Instead, you get a sliding scale of bonus points for beating teams in the Top 20 of the RPI. That means your bonus points can change as teams move around or drop in and out of the Top 20, but it won't change drastically because it's a sliding scale.
So, expect to see less volatility as selection day approaches, which is precisely one of the things the Committee wanted to accomplish with the changes.
* * *
Taking a look at the "Weighted Win Percentage" on our Pairwise chart, you can see the teams being helped and hurt the most by the disparity in their home/road record. While "Weighted Win Percentage" is not directly calculated into the RPI, it is indirectly, via the various RPI components. In the Top 20, the teams most negatively affected are Notre Dame and Wisconsin. The teams most positively affected are Boston College and Ferris State.
Another thing that changes this year, although not because of a change to the formula — with six conferences now, the odds of having to swap teams to avoid first-round intraconference matchups goes down. But we'll deal with that as we get closer to seeding.
1-3. Minnesota, Boston College, Quinnipiac
At this point, these are very likely to be No. 1 seeds. You can kind of see this rounding into form already. Minnesota would, of course, be in St. Paul; BC in Worcester and Quinnipiac in Bridgeport. Works out rather tidily. There will be a big difference, however, between the top seed and the second overall seed. No. 2 would play 15, and that's currently New Hampshire. The No. 1 overall seed will get to play the eventual Atlantic Hockey champion.
Which brings us to this point: No Atlantic team is going to be winning at-large bids this year, that's for sure. Last season, there was almost two that did so. This year, the highest-rated AHA team currently is Mercyhurst at 37, with an RPI below .500.
4-6. Union, Ferris State, Cornell
These teams are bunched up in another grouping, and if it stays similar, then they'll be battling it out for a No. 1 seed in the Cincinnati regional.
In this grouping, we see a direct affect of the new system, and how a team's record against "good teams" can actually carry more weight than in year's past, even though there is no "Record vs. TUC" criterion.
Union has a good-sized Quality Win Bonus (QWB) by virtue of some good wins, while Ferris State has very little. As a result, Union surpasses Ferris State in RPI and wins the comparison even though FSU clobbers Union in Record vs. Common Opponents.
In the past, Ferris State would've won Record vs. COP and RPI, while Union would've won Record vs. TUC — giving FSU that comparison. By taking Record vs. TUC out of the equation and folding it into the RPI, you give a very powerful boost to a team for "good wins," yet you're doing it in a non-volatile way.
7-10. St. Cloud State, Lowell, Michigan, Northeastern
This is the next bunching of teams, and if things were seeded today, would play against each other. Because of the reduced volatility, these teams can already start feeling relatively comfortable for making the NCAAs, even though there's a ways to go.
Another example here of how RPI means so much. Providence has a better Record vs. COP than Northeastern, but is behind in RPI and lost a game to Northeastern this season. Providence could sweep Northeastern in a playoff series and flip the comparison, but by that point, its RPI will probably be better anyway.
12-14. Minnesota-Duluth, Wisconsin, North Dakota
Wisconsin is interesting in this group because it's the first team where we can clearly see that an inability to win on the road is hurting its RPI. As a result, Wisconsin is more of a bubble team than it otherwise would be, thanks to the new formula.
15-16. New Hampshire, Notre Dame
If things ended today, New Hampshire would get in the NCAAs and Notre Dame would not, thanks to UNH's sweep of the Irish last weekend. UNH is the first team on the Pairwise list that overtakes another team despite having a lower RPI, based on the other criteria. If this holds up, it would also be the first team to have to swap places with another team to avoid a first-round matchup against a team from its own conference (Boston College). It would be an easy swap with the 14th team.
Notre Dame, like Wisconsin, is getting hurt by its weak road record.
17-19. Clarkson, Vermont, Maine
Clarkson loses a comparison to Colgate, because Colgate has won two games head to head this year. That is holding Clarkson down a bit and will hurt its chances. This can be overcome if Clarkson get beat out Colgate in Record vs. COP, which is close.
20-21. Yale, Colgate
While there is no "TUC Cliff" anymore, there is still a soft "TUC Ledge." And the 20th spot is it. If you've beaten Yale this year, you get a small RPI boost. If you beat Colgate, you want the Raiders to move into 20th so you can get the benefit of the boost.
Colgate also has the third-highest QWB of any team in the country, meaning it has beaten a number of highly-rated teams.