Men's at-large berth analysis and predictions

By Christan Shirk

PART I

The conference tournaments will wrap up Saturday and Sunday and 42 of the 61 berths in the NCAA men's tournament will have been claimed via automatic qualification (AQ). That leaves 19 at-large berths for the men's committee to award by selecting one team from Pool B and 19 from Pool C. (For a full explanation of the Pool B and Pool C classifications and the allocation of berths see the column AQ's, Pool B and Pool C? What does it all mean?) So who's in the running for the 18 Pool C and 1 Pool B at-large berths? Who will be dancing and who's season has come to a close?

Selection Criteria

Well, first let's quickly review the criteria the committee uses to evaluate teams and make the at-large selections. The selection criteria are found in Section 2.4 of the Pre-Championships Manual (pg. 21) and are divided between primary criteria which consider each team's results against other Division III opponents and secondary criteria that introduce matches against non-Division III opponents (NAIA, NCCAA, Division II, Division I), the later only being considered if the former does not enable a distinction to be made between schools. The criteria underwent a significant change prior to last season (2013) when the in-region/out-of-region distinction was abandoned. Previously only games versus "in-region" opponents were consider as part of the primary criteria with secondary criteria considering out-of-region and non-Division III competition.

Beyond the selection criteria, note the following principles concerning at-large selections.  

• Teams are selected on a national basis, using in-region selection criteria.
• There will be be no predetermined regional allocations for Pools B and C.
• There will be no maximum or minimum number of berths from one region.

Background: Observations from the Past

I will use the third and final NCAA weekly regional rankings as the starting point to evaluate the Pool B and Pool C landscape. The reason for this is that these rankings are done by the same national and regional committees which will make the at-large selections and are done by applying the same criteria which is used for making the at-large selections. Therefore, by design, the NCAA regional rankings are a direct foreshadowing of the at-large selections providing a certain level of transparency to the at-large selection process. It is for this reason that these rankings are so important and insightful.

Furthermore, a comparison of the at-large selections and regional rankings over the past several years yields the following observations.  

• In the past seven years (2007-2013) no team that was unranked in the third weekly regional rankings (those released the Wednesday before the selections) was selected.
• There will probably be twice as many Pool C candidates in the rankings as available berths.  For example, last year there were 43 Pool C teams in the final rankings but only 19 Pool C berths available.  In 2012 the ratio was 38/20 and in 2011, 40/19.
• Regions generally have two or three ranked teams not selected, maybe one more or one less for the weakest and strongest regions in that particular year.
• Within a region, rarely does a lower ranked team in the third weekly rankings get selected ahead of a higher ranked team. For example, there was only one instance of this last year out of the twenty men's at-large selections.  In 2012 there were two cases while 2011 represented a high of three instances and 2009 had none.  The few times a team has been selected instead of another that had been ranked higher, the results after the final rankings served as a very reasonable explanation.
• Most of the at-large selections come from the top half of each region's rankings.
• There has been no indication that that conference tournament results are weighted extra because they are the most recent results nor because they may be considered "big" games, and they shouldn't as the primary and secondary selection criteria makes no such allowance.

This leads to the following conclusions:  

• A team that is not ranked in the third and final regional rankings has virtually no shot at a Pool C berth.
• Many ranked teams will not be participating in the tournament.  It isn't good enough to simply be ranked to gain an at-large berth.
• In fact, a team usually needs to be in the top half to two-thirds of their regional rankings to be selected for an at-large berth.
• Do not expect big jumps or falls due to the final week's results (mostly conference tournaments), which makes sense as one week only represents about 10% of the total schedule and conference tournament results are not weighted extra. Furthermore, any team in need of an at-large berth presumably lost or tied in the final week minimizing chances they would climb the rankings.

PART II

Using the Regional Rankings as a Guide

Well, that was quite the introduction. Let's now take a look at the third week of the NCAA regional rankings to see where things stand. I have listed the rankings and color-coded the teams according to their Pool:  Pool A (AQ), Pool B, and Pool C. A couple of notes to properly understand the data in the table: the Strength-of-Schedule (SOS) value is from last week's rankings and I will not be trying to update it for this past week's results. However, I have updated the other information (overall record, Division III record, record versus ranked teams). I list a team's ranking in each of the three weekly releases with the order based on the third ranking.

PART III

Pool B Analysis and Predictions

There is just one Pool B berth available for the three Pool B teams in the third regional rankings: UW-Whitewater, UW-Oshkosh, and UC Santa Cruz. For ease of comparison, let's put the three Pool B teams in their own table.

I don't think UC Santa Cruz quite measures up to the two Wisconsin schools, but the independent school isn't as far behind as many may think because a number of Santa Cruz's losses came against non-Division III opponents which doesn't get considered in the primary criteria. And their SOS is actually quite good and higher than the two Wisconsin schools and they do have a win over ranked North Park on their resume. But I have to think the committee recognizes that the West region isn't as strong as the North region and that keeps UC Santa Cruz from staying in the conversation. The committee will probably be struggling to pick between Oshkosh and Whitewater. The gap in rankings (No. 3 to No. 7) would usually be just too much for the last week's results to close, and since Whitewater didn't drop two games one should probably err on the side of Whitewater in making a prediction. But, they went head-to-head twice, and I think that Oshkosh's win in the WIAC final (after a regular season tie) could tip the scales in their favor if otherwise they are considered even. But are they? Same winning percentage, though Whitewater with one more blemish and Oshkosh with one more loss. SOS was very close last week, and should have nearly evened out with Whitewater's WIAC semifinal opponent having a better winning pct. than Oshkosh's. Whitewater beat both North Park and Milwaukee Engineering by 1-0 scores while Oshkosh lost 2-3 and 0-2, respectively. However, Oshkosh tied Loras 2-2 while Whitewater lost 0-3 on opening day. And Oshkosh beat Concordia (Wis.) 4-1; Whitewater only tied 1-1. Their updated records versus ranked teams favors Whitewater (2-2-1) over Oshkosh (1-2-2), that owing to Oshkosh losing 0-1 to MIAC champion Gustavus Adolphus and Whitewater defeating Chicago, champion of the tough UAA, 1-0. I think that's the difference in UW-Whitewater holding off Oshkosh to gain the Pool B berth despite the head-to-head loss, but I also wouldn't be completely surprised if the nod went to UW-Oshkosh on the strength of the head-to-head win. In any case, the team passed over could still be in the running for a berth for Pool C.

The Pool C Candidates

After an unusually high number of Pool C teams in the third regional rankings last year (44), this year sees the number return to the norm. In 2014 the final released regional rankings contain 36 Pool C teams, exactly double the 18 available Pool C berths. So, which 18 of the 36 teams in blue will get selected by the committee? Who will be left out? Well, actually, it's more than 36 as the two Pool B teams not awarded the single Pool B berth will land in Pool C. So make that 38 ranked teams fighting for 18 Pool C berths.

Let's start by throwing all 36 of the ranked Pool C teams together in one table along with the three ranked Pool B teams. The eighteen at-large selections will come from this list.

Pool C Analysis and Predictions

POOL C LOCKS (4)

1. Brandeis (17-2-0) - Extremely high SOS (highest in nation), .895 win pct. (sixth best in nation), and superb 7-2-0 vs. ranked teams. Were first in New England rankings all three weeks.

2. Franklin and Marshall (15-1-2) - High SOS, solid 3-1-1 vs. ranked teams, and were one of only three teams to go undefeated in the regular season despite playing in the tough Centennial Conference. Were second in the Mid-Atlantic rankings behind Messiah all three weeks and their surprise conference semifinal loss will only see them slip down to third, if that.

3. Loras (15-2-2) - High SOS, great 5-0-2 vs. ranked teams, and win pct. approaching .850. Were first in North rankings all three weeks

4. Emory (13-3-2) - Very high SOS, impressive 5-2-2 vs. ranked teams, and win pct. over .750. Were first in South Atlantic rankings all three weeks.

SAFE, NO WORRIES (4)

5. Wheaton (Mass.) (16-3-2) - Very high SOS, win pct. over .800 with acceptable 2-2-2 record vs. ranked teams. Were ranked second in New England rankings last week.

6. Dickinson (11-5-2) - Very high SOS with 3 wins versus ranked teams compensating for win pct. below .700. Achieved a good 3-3-1 record vs. ranked teams, and are best in the Mid-Atlantic after the obvious top three who have a combined 2 losses among them.

7. Ohio Wesleyan (15-4-2) - Decent SOS, win pct. over .750 and solid 3-2-2 record vs. ranked teams. Firmly ranked third in Great Lakes all three weeks and perhaps moved ahead of John Carroll this past week into second.

8. Cortland State (14-4-1) - Win pct. over .750 and good 3-2-1 record vs. ranked teams compensates for so-so SOS. Tied and narrowly lost to undefeated Oneonta St. and were firmly third in the East rankings all three weeks

IN GOOD SHAPE (4)

9. Coast Guard (13-2-3) - High SOS, win pct. over .800 with acceptable 2-1-1 record vs. ranked teams, were ranked 2nd-2nd-3rd in New England rankings and conference semifinal loss to WPI shouldn't drop them further than 5th if that and the New England region always gets their top six teams in at a minimum.

10. Rutgers-Newark (15-5-0) - Their .750 win pct.matches the best of the ranked teams still on the table and no other remaining candidates have more then two wins versus ranked teams. So an adequate 2-2-0 record vs. ranked teams combined with a decent SOS should kept them from being among that last couple teams selected. They were third in last week's South Atlantic rankings and despite their conference semifinal loss it's possible they pass Salisbury whose own conference semifinal loss took the shine off thier to-then one-loss season and left them without any wins versus ranked opponents.

11. North Park (13-5-1) - Surprisingly missed out last year, being passed over for Misericordia. All numbers (win pct., SOS, vs. ranked) are a little but not much better this term, and sitting in 4th last week versus 2nd a year ago has got to leave the team very nervous. But the Pool C landscape is different and it's hard to see the committee snubbing them again with their very high SOS and a second win versus ranked this time around.

12. Amherst (13-1-4) - Usually a team needs a win versus a ranked team to get an invite. No team without one got selected in the past three years, and here sits Amherst, the highest ranked team from arguably the toughest conference, without one. Their SOS is high to very high and they only suffered one loss for a win pct. over .800. Whatever had them in front of Tufts in last week's rankings should keep them there again, so they'll be at 4th or 5th (Babson could pass them, but Coast Guard could fall behind them) in New England. The NESCAC is not going to be shutout of any at-large berths and New England is always going to get at least their top six teams in, so the committee looks the other way on one of their favorite criteria: wins vs. ranked teams.

ON THE RIGHT SIDE OF THE "BUBBLE" (3)

13. Tufts (10-2-4) - High SOS and their .750 win pct. matches the best of the ranked teams still on the table. 1-1-2 record versus ranked teams isn't impressive, but just enough given their high SOS. After week off due to NESCAC quarterfinal upset loss, should remain sixth in the New England rankings ahead of WPI which has always been good enough to get selected.

14. Rochester (9-5-3) - Five losses is pushing it, but not uncommon. Eight blemishes usually is too much though Oneonta got in last year with eight (10-4-4). But an extremely high SOS (second highest in nation) and four wins vs. ranked teams (4-3-1) is what the committee loves to reward. Carthage missed out last year with a similar 4-3-0 record vs. ranked teams, but did not have nearly as high a SOS (.566) nor the history Rochester has of being shown favor by the selection committee.

15. John Carroll (17-4-0) - Win pct. over .800 and a pair of wins versus ranked opposition is just enough to makes their low SOS passable. Firmly held second place in all three Great Lakes rankings, and while highly likely they get paased by Ohio Wesleyan, third is all the further they should fall.

ON THE "BUBBLE" - PICK 'EM (3 of 6) (listed in alphabetical order)

Brockport State (10-3-6) - No Pool C team has been selected with nine blemishes in the past nine seasons. But the committee continued to rank them fifth in the East region despite the blemishes and there's little reason to think they'd be passed by RPI or anyone else during the past week. Their SOS is borderline high, but they only have a single win versus a ranked opponent.

Haverford (11-7-1) - Very high SOS and three wins versus ranked teams, but their seven loses is more than any Pool B at-large selected team has carried into the tournament is at least the last nine years. However, this year's field of Pool C candidates makes it a possibility evidenced by the committee ranking them fifth in the Mid-Atlantic last week depsite 6 losses and only a single win versus a ranked opponent. They've since added two wins versus ranked teams, including highly ranked and previously undefeated F&M, in the final week before being hit with the seventh defeat. Taking a fourth team from the same conference might be against them, though that's never been a problem for the UAA.

Salisbury (12-2-5) - They worked themselves up to second in the South Atlantic by the final rankings, but then lost in the CAC semifinals. Now instead of no wins or losses against ranked opposition, they are just missing a win versus a ranked team. As said above, in the past three years no at-large selection has been without one and all other suspected "bubble" teams have at least one. They do hold the highest winning pct. (.750) among the suspected "bubble" teams the second lowest SOS of the group. Given how Montclair State went from second to seventh due to a single loss versus a ranked opponent, it's difficult to predict where the committee will move Salisbury—holding at second or falling to fifth wouldn't be a surprise which is fitting for a tram that in recent years has been both a surprise inclusion and surprise snub.

UC Santa Cruz (11-6-1) - Their high to very high SOS is what may get this team from Pool B into the discussion after having spent the past two weeks ranked fourth in the West region. Misericorida got selected last year with five losses, a lower SOS, and just one win versus a ranked team, so Santa Cruz may be viable. A sense that the West region is relatively weak may be held against them and it should be remembered that no at-large berth went to the region last year and just one the year before.

UW-Oshkosh (14-5-2) - Probably edged by UW-Whitewater for the Pool B berth despite their WIAC final head-to-head win over their rvials, but still in the running for a Pool C berth due to decent SOS, a win pct. over .700, and a win over a ranked opponent. The pair of WIAC tournament wins should see them pass Carleton and St. Olaf in the North region. Their resume is not out of place among the suspected "bubble" teams and their 1-2-2 record versus ranked teams isn't clearly bettered by any of the others.

WPI (14-5-0) - Second highest winning pct. among the suspected "bubble" teams and their two wins versus ranked teams is more than the others except for Haverford. But they have a fairly low SOS of .533 which would be the lowest among the predicted at-large selections. They were seventh in the New England rankings last week and In many past seasons that would probably see them getting an invite, but the strength of the region has been suspect even if the increased parity is not.

A farily decent case could be made for unranked Kalamazoo making onto the "bubble", but if the committee was going to give them serious consideration, it wouldn't make sense for them to have not been included in last weeks rankings. Until a team that was unranked the previous week actually gets selected (it's been eight years since it's last happened) the assumption will remain that that would be too far of a jump.

Forced to forecast the committee's final three selections, I'll go with Haverford, UW-OShkosh, and Salisbury.

Some Final Observations and Comments

● First, these are my predictions of who the committee will select, not who I think they should select or who I would select.

● This year there were two cases in which the outcome of a conference tournament clearly cost the rest of Pool C a berth to the tournament. First is Bowdoin taking the NESCAC automatic berth instead of Tufts or Amherst. And there was Heidelberg's surprise win over #1 seed John Carroll in the OAC final. Babson beating Wheaton (Mass.) in the NEWMAC semifinals may also have taken a Pool C berth off the table, but only if Babson wasn't going to get selected with a sixth loss coming in that semifinal instead of a win. F&M 's upset loss only hurt other Pool C candidates if that wins is what earns Haverford an at-large berth as the Centennial was probably always going to get three teams in, in which case who takes the automatic berth doesn't really effect the rest of Pool C. That's just enough that several "bubbles" teams that don't get invited will wonder if that's what cost them, but it cold have been much worse and there's been years where it has been.

● Before delving into the data and comparisons, I expected it to be more difficult than ever to make predictions. But to my surprise I didn't struggle nearly as much as anticipated to come up with my list. Of course, my predictions could prove to be way off (or at least less accurate than in previous years), but by trying to follow the approach I've taken in previous seasons—that is, to assume consistency between the rankings and the selections, to not over-think things and go with what seems obvious, and never get distracted by my personal opinion on who should or shouldn't be selected—I didn't find myself second guessing myself and re-doing my picks much at all.

● That said, we saw some isolated volatility in the rankings and more head-scratchers than usual this year, so in that sense if makes me less confident in my predictions because the committee has seemed to become a little less predictable.

● The possibility of a 7-loss team and/or a 9-blemish team getting selected makes this year more interesting. Also, interesting is the real possibility that a NESCAC isn't the most represented conference from the New England region and that the Centennial could get four team in, something usually reserved for the UAA and maybe the NESCAC. And Salisbury seems to regularly position themselves to be a questionable selection or omission, helped this year by a South Atlantic region that wasn't always easy to figure out nor the NCAA ranking of their teams.

● Well, I honestly don't feel especially qualified to make these predictions each year, but I appreciate all the interest this column continues to generate each year. In any case, hopefully it's always informative, interesting, and fun for you, the readers, regardless of the accuracy of my picks.

Comments or feedback for the author?  Email Christan Shirk.