<
>

Geography, RPI not clear-cut

This year's bracket offered few surprises, but was very revealing at the same time. From seeds to the at-large teams selected, the bracket shaped up largely as expected, but there's still plenty to analyze.

Geography didn't matter at all

When it came to the top seeds at least, the committee stuck strictly to the S-curve and didn't apply those rankings geographically. The surprise is that it goes against one of the NCAA's stated goals. As stated in the principles and procedures for bracketing the teams: The committee will attempt to assign each team to the most geographically compatible regional and first‐/second‐round site, by order of the S‐curve

The committee doesn't reveal its 1-64 list of teams, so the cloak and dagger around this will remain, but it would be difficult to convince anyone that Maryland wasn't No. 4 overall, Tennessee No. 5 and Kentucky No. 8. If this was done geographically by S-curve, as has been the way in recent years, the Lady Vols should have been in Albany, New York, with No. 1 overall seed Connecticut. Instead, Tennessee was placed with the Terrapins, the No. 4 and No. 5 together, just like old-school bracketing and the way fans and media tend to like better.

However, this still represents a year-to-year inconsistency in how the bracketing is done. One season the so-called G-curve is applied; in others, like this one, it is not.

South Carolina, as the top seed in Greensboro, North Carolina, is the sensible thing to do, but it too goes against the geographic principle if we assume Notre Dame is No. 2 overall and the Gamecocks are No. 3. That one is a little easier for the committee to get around because those teams are so close. Putting South Carolina closer to its home also allowed the committee to get another ACC team (North Carolina) in Greensboro. From an attendance and fan perspective, that is a good thing.

RPI is just a guide

This has been written about countless times, yet a misconception still exists as it pertains to the relationship between RPI and seeding. The committee uses RPI as merely a guide to group and begin the comparison of teams. This bracket, more than ever, proves there simply isn't a direct correlation between the two. The examples are many this year:

• George Washington: RPI 11, No. 6 seed
• Princeton: RPI 12, No. 8 seed
• Dayton: RPI 13, No. 7 seed
• FGCU: RPI 14, No. 7 seed
• Chattanooga: RPI 20, No. 7 seed
• Western Kentucky: RPI 27, No. 12 seed
• UALR: RPI 29, No. 11 seed

Let's zero in on Princeton. With the Tigers' perfect 30-0 record, many people believed they deserved as high as a top-four seed, but their RPI and seed represent the biggest discrepancy between the two.

The committee, as it routinely maintains, puts a number of factors into perspective. This isn't to say RPI means nothing, but this illustrates once again that it isn't everything. Yes, Princeton was 30-0, but the committee also recognized that 21 of those wins were against opponents from outside the RPI's top 100, that the Tigers beat only one tournament team, and that their strength of schedule and nonconference SOS were both rated outside the top 100. In mid-February when the committee released its top 20 teams, Princeton didn't make the cut. That should make Princeton's No. 8 seed Monday a little less surprising. Should an unbeaten team, regardless of strength of schedule, be seeded higher than a No. 8? Perhaps. But it's just another example of many that RPI should not be viewed so strictly.

Princeton, however, was put in a difficult spot by the committee and not at all rewarded for its 30-0 record. The first round pits the Tigers against another strong mid-major in Green Bay, with No. 1 seed Maryland likely awaiting them in the second round. Princeton's stated goal of winning at least one tournament game will be tough enough. Getting to a Sweet 16 now appears virtually impossible.

George Washington out as host

The Colonials are another one of those teams for which RPI and seed were not aligned, but the stakes might have been even higher for GW. The Colonials were in the committee's top 20 in mid-February, meaning they were at worst a No. 5 seed on Feb. 11; they were in a position to possibly host the opening two rounds.

Instead, George Washington lost -- on that very same night -- to a sub-100 St. Louis team. The Colonials didn't lose again and went on to win the Atlantic 10 regular-season title and tournament crown. No matter. That one loss seemed to have cost George Washington the chance to host, as well as at least one seed line.

Same conference, same side of the bracket ... didn't matter

When it came to separating conference foes from possibly meeting before the regional final, the committee could have done better. Teams from the same conference are placed in the same half of a regional bracket four different times in this bracket:

• Syracuse and North Carolina in the upper half of Greensboro
• Nebraska and Ohio State in the upper half of Greensboro
• LSU and Kentucky in the upper half of Albany
• Iowa and Northwestern in the lower half of Oklahoma City

Technically, the committee does not have to avoid conference foes meeting until the regional final, but it is supposed to try, as written in the principles and procedures: The committee will attempt to keep conference teams from meeting until the regional final round.

Strictly speaking, the only time this is firm is if the teams have played three or more times in the regular season and conference tournament. However, allowing this in the bracket in four different places isn't putting a whole lot of emphasis on the "will attempt" part.

Two of these instances could have been avoided simply by switching the regions of teams on the same seed line. It would seem hard to rationalize that, for example, there is such a difference in Dayton and Northwestern that if those two switched regions it would gravely mess with the balance of the bracket. It wouldn't. However, it would prevent Iowa and the Wildcats from potentially meeting for a third time this season in the regional semifinals.

By no means does this ruin or hurt the bracket in any significant way. But if the stated rules say conference foes meeting before the regional final is something the committee would like to avoid, shouldn't this committee have at least tried to do so? Four different occurrences doesn't look like it did.

Cats turn nine lives into a No. 2

Kentucky scheduled tough and clearly got rewarded for it. Nine losses did not seem to bother the committee. It was the strength of schedule that mattered more. The most losses for any No. 2 seed last year was six. The Wildcats were good against a difficult slate with wins over Baylor, Louisville and South Carolina, but they were also blown out by Duke and Texas A&M and lost to sub-100 Illinois and Ole Miss.

This isn't unheard of, but it's rare for a team with so many losses to be seeded so high. Tennessee was beaten nine times in 2012 and was a No. 2. The commonality? Schedule strength. The SOS of both was rated in the top five. The moral of the story seems to be schedule hard if you can, win a couple of big ones, and good seeding will come your way.

Last place is good here

The last four teams in the field were LSU, Miami, Oklahoma State and Tulane. With Arkansas' inclusion as a No. 10 seed, it also means all 32 at-large berths were correctly predicted here.

What's interesting is that the 6-10 SEC team in Arkansas had more of a cushion for being included and was a seed line higher than 10-6 LSU. The Razorbacks' overall profile might have been slightly better with out-of-conference wins over Iowa and Oklahoma, and better RPI and SOS, and that is exactly what the committee was telling us with this seeding comparison. Certainly, conference record was part of the mix, but it didn't win out in this case. Arkansas' convincing win over the Lady Tigers probably contributed as well.