Men's at-large berth analysis and predictions

PART I

By Christan Shirk

The conference tournaments will wrap up Saturday and Sunday and 42 of the berths in the NCAA men's tournament will have been claimed via automatic qualification (AQ). With a full 64-team tournament field this year, up two from previous seasons, that leaves the men's committee with two more at-large berths to award this year—22 total—by selecting one team from Pool B and 21 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 21 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 2019 Division III Soccer Pre-Championships Manual (pg. 23). The criteria is divided between primary and secondary criteria, the latter only being considered if the former does not enable a distinction to be made between schools. Regular season and conference postseason matches are considered.

For further explanation of the criteria, including Strength of Schedule, results versus ranked teams, and a list of secondary criteria, go here or here.

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.

Selection Committee

The NCAA championship tournament is administrated by the eight-member NCAA Division III Men's Soccer Committee which is composed of the chairs of their respective regional advisory committees. These committees make the at-large selections to complete the tournament field, assisted in the evaluation of teams by the Regional Advisory Committees. These are the same national and regional committees which release the pre-tournament weekly NCAA Regional Rankings. The members of these committees can be found on pages 9-13 of the Pre-Championships Manual.

At-Large Selection Timeline and Process

The process of making the at-large berth selections for the NCAA tournament starts with weekly NCAA regional rankings mentioned above. These rankings are done following the fourth last, third last, second last and last week prior to the tournament selections being made. The first three of these weekly rankings are, by design, a direct foreshadowing of the at-large selections because they are (1) done by the same committee that makes the at-large tournament selections and (2) done by applying the at-large selection criteria. You can learn more about the NCAA Regional Rankings here or here.

Following the release of the third weekly regional rankings the process is as follows.

• Conference championships are completed by 6:00 p.m. ET, Sunday, November 10.
• The NCAA compiles the data corresponding to the at-large selection criteria (win-loss-tie percentage against Division III opponents, results versus ranked Division III teams, Division III Strength-of-schedule) and provides it to Regional Advisory Committees.
• The Regional Advisory Committees do their fourth regional rankings in the same manner as the previous three weeks. The results versus ranked Division III teams (RvR) criteria is based on who was ranked in the third regional rankings that were released on Wednesday, October 31.
• The national committee makes adjustments to the regional rankings as they see fit but does not publish them until after they have announced the tournament field (including the at-large berth selections).
• An updated RvR is developed based on opponents were ranked in either the third or the just completed fourth regional rankings. This is the RvR that the national committee will use when comparing teams across regions on a national basis.
• Pool B teams (independent institutions and institutions that are members of conferences that do not receive an automatic berth in the tournament) in the final regional rankings are identified.
• The highest ranked Pool B candidate from each region is placed "on the board", the teams are discussed, and one team is selected for the lone Pool B berth.
• Pool C teams (teams who were not awarded their conference's automatic berth and unselected Pool B teams) in the final regional rankings are identified.
• The highest ranked Pool C candidate from each region is placed "on the board", the eight teams discussed, and one team is selected. The next highest ranked Pool C candidate from the selected team's region is added to the board and the process repeats until all 21 Pool C at-large berths have been awarded.

The 22 at-large selections are added to the 42 teams who were awarded their conference's automatic berth, completing the 64-team field at which point the committee begins the process of grouping the teams and developing the tournament bracket with geographical proximity playing a major role. You can read more about that here.

Background: Observations from the Past

The third NCAA weekly regional rankings, the last rankings published prior to the tournament field being announced, will be used 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 make the at-large selections and are done by applying the same criteria that 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. (Note: A fourth ranking is done Sunday night after all conference championships are complete and serves as the basis for the at-large selections, but those rankings will only be published after the tournament field is announced.)

Furthermore, a comparison of the at-large selections and third weekly regional rankings over the past several years yields the following observations (for an expanded discussion of these observations, go here).  

• In the past twelve years (2007-2018) only once—just last year—has a men's team that was unranked in the third weekly regional rankings (those released the Wednesday before the selections) been selected. This might occur more often moving forward if the increased parity currently being seen continues combined with this year's addition of two more at-large berths.
• Even with two more at-large berths available this year, there will probably be about twice as many Pool C candidates in the rankings as available berths. For example, last year there were 46 Pool C men's teams in the third rankings but only 19 Pool C berths available.  In the five years before that, the ratio was 42/19, 41/19, 39/18, 38/18 and 44/19.
• 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.
• The ranking of teams changes little in the last week and consequently, within a region, rarely does a lower ranked team in the third weekly rankings get selected ahead of a higher ranked team. That said, last year, for the second year in a row, there were three instances of this happening which exceeds the one or two occurrences that had typically been seen in previous years. Again, this might continue to occur more frequently due to increased parity and the increased number of at-large berths.
• Most of the at-large selections come from the top half of each region's rankings.
• There has been no indication 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 regional rankings has very slim chances of receiveing 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

By Christan Shirk

Using the Regional Rankings as a Guide

Well, that was quite the introduction. Let's now take a look at the teams that were ranked this year as well as a few others, especially any teams that we think may be ranked in the fourth regional rankings. We'll list the teams as ranked in the third published rankings followed by previously ranked teams and the additional teams. The teams are color-coded according to their Pool: Pool A (AQ), Pool B, and Pool C. The Division III record, winning percentage, and record versus ranked teams have been updated to account for the results this past week that were not accounted for in the third weekly rankings. The Strength-of-Schedule (SoS) value is an approximate updated value that should be within a several percentage points.

The Pool B Candidates

The third published NCAA regional rankings did not contain a single Pool B team. UC Santa Cruz was ranked in the first weekly ranking for the West Region. So let's just show all the Pool B teams with at least a .500 winning percentage for comparison purposes.

The Pool C Candidates

The third published NCAA regional rankings contained 45 Pool C teams, which is one less than last year but still higher than typical. More importantly, it is more than double the 21 available Pool C berths. To those 45 we will add a few other Pool C teams that could be ranked in the fourth rankings. So, which 21 of the ranked or potentially ranked teams in blue above will the NCAA committee select? Let's start by grouping those Pool C candidates in the table below. The 21 at-large selections will come from this list.

PART III 

By D3soccer.com Contributer

 Pool B At-Large Berth Analysis and Predictions

Looking at the table of Pool B candidates in the previous section, it seems that UC Santa Cruz (8-7-4 overall, 5-2-4 against Div. III opponents), UW-Platteville (8-8-3) and UW-Whitewater (15-4-2) are the leading candidates. Platteville has a strength of schedule on par with typical Pool C selections (0.573) while Whitewater has a mark (0.500) that would keep them far from consideration in Pool C. But Whitewater's winning percentage (0.762) is far superior to Platteville's even record (0.500). Santa Cruz, who was ranked fifth in the week one West Region rankings, falls in between on both these criteria. Despite the Banana Slugs tying Whitewater in the ACAA final (the Warhawks took the title on PKs), the berth likely comes down to the pair of Wisconsin state schools who both have a win versus a ranked team: Platteville beat St. Norbert (to whom Whitewater lost) and Whitewater edged Loras (who beat Platteville). Both lost to highly ranked Luther and defeated unranked Edgewood and Wartburg. This selection could go either way, but because the Warhawks walked away from their head-to-head encounter at Platteville with a 1-0 win, we'll tip Whitewater for a place in the NCAA tournament.

 Pool C At-Large Berth Analysis and Predictions

Before making our predictions, a few insights into the process to narrow down the selections. First, we picked teams in the order we anticipate the committee will select them. Second, once we picked a team, we removed them from discussion, so we compared only the remaining Pool C teams. Third, given the committee’s long history of not putting unranked Pool C teams into the tournament, we limited consideration to those teams who appeared in last week’s regional rankings—45 teams for 21 spots. Finally, some of those teams—such as those with zero ranked wins—have almost no chance of making the field. So while it’s possible that teams such as Elizabethtown or Redlands could make the field, we did not consider them realistic possibilities for a bid. (The one exception is St. Thomas—more on that below.)

LOCKS - NO DISCUSSION NEEDED (3)

1. Amherst (14-1-2) - Probably the first name off the board. The Mammoths have held the top spot in each edition of the New England rankings, and a narrow 1-0 loss to Middlebury—their only of the season—won’t change that. A near-.900 winning percentage (highest in Pool C) and stellar 4-1-2 mark against ranked teams makes this an easy choice.

2. Franklin and Marshall (16-3-0) - Our other candidate for the first at-large selection, the Diplomats check every box: a top strength of schedule (over .600), a great winning percentage, and six ranked wins—tied for the most in Pool C.

3. RPI (14-2-3) - The strength of schedule is average, but with a winning percentage over .800 and six ranked wins—tied for the most of any remaining Pool C team—RPI will be in very quickly.

SAFE, NO WORRIES (5)

4. Connecticut College (10-3-3) - The winning percentage is above-average for the remaining teams, the strength-of-schedule is over .600, and three ranked wins (and a .500 record-versus-ranked) gets the Camels over the line. Middlebury could jump the Camels, having beaten Amherst in the NESCAC tournament, but Connecticut College’s higher winning percentage and better record versus ranked teams makes them a slightly safer play.

5. Christopher Newport (13-2-4) - The Captains benefit by being on the board from the start, as they are the top Pool C team in the South Atlantic. Newport only has two ranked wins, which is near the bottom of the realistic Pool C candidates. But the Captains went .500 in seven games against ranked teams, which is more than enough of a track record for the committee. Add in an excellent strength of schedule (top 10 of those remaining) and one of the highest remaining win percentages, and you have an easy pick.

6. Ohio Wesleyan (13-5-3) - The Battling Bishops’ profile tells a similar tale. The winning percentage is not great, especially in isolation, but it’s in the top half of the remaining Pool C candidates. Mix in the same strength-of-schedule as Christopher Newport and five ranked wins—the most of any team remaining in Pool C—and OWU is safe.

7. Central (14-3-2) - The strength of schedule is on the low side, even of those teams remaining, but the Dutch have one of the best winning percentages of the realistic remaining Pool C candidates and a strong 3-1-1 record against ranked teams. The North region is down overall this year, but Central is in.

8. Ohio Northern (13-6-2) - Another similar profile, the Polar Bears have one of the highest remaining strengths of schedule and four ranked wins (tied for the most remaining). A .500 record in nine games against ranked teams will go a long way with the committee.

IN GOOD SHAPE (3)

9. Middlebury (8-3-7) - The only worry for Middlebury is that it is one of the only Pool C teams with more blemishes than wins. In other words, Middlebury won fewer than half its games. In many years that would sound a death knell to any tournament hopes. But Middlebury’s strength of schedule is a sky-high 0.629. The winning percentage is in the bottom half of the remaining, realistic candidates, but it’s certainly not disqualifying. So between the strength of schedule and the two ranked wins—including a season-saving victory at Amherst in the NESCAC tournament—Middlebury should be just fine.

10. Rowan (11-5-3) - Rowan should jump Ramapo in the South Atlantic after avenging a regular-season loss with a 1-0 win in the NJAC tournament. That win also counted as the Profs’ third versus ranked teams, giving it by far the best winning percentage against ranked opponents of the remaining teams near the bubble. Throw in a 0.620 strength of schedule, and Rowan should make the field comfortably.

11. Rochester (11-5-1) - Ranked wins to the rescue. Rochester’s profile would push it much closer to the bubble in many years, given the average strength of schedule and winning percentage. But a 4-3-0 record versus ranked teams is good in any season, and it should be enough here. We expect Rochester to jump Ithaca, given Ithaca’s loss last week and Rochester’s better record versus common opponents.

NERVOUS, BUT PROBABLY OKAY (4)

12. Swarthmore (11-3-3) - Two ranked wins is not a great stat, but Swarthmore’s overall profile is balanced compared with those teams remaining. It has one of the best winning percentages and one of the best strengths of schedule left, and two ranked wins are about average compared to those teams remaining. As the committee gets deeper into the selection process, that balance will put the Garnet in a stronger position than other remaining teams.

13. Williams (8-4-5) - Like NESCAC counterpart Middlebury, the Ephs are the rare team with more blemishes (nine) than wins (eight). The difference is that Williams has double the losses and a weaker strength of schedule. But that “weakness” is relative, as Williams still has one of the strongest remaining strengths of schedule plus three ranked wins. While we expect Middlebury to jump the Ephs, they are in fairly good shape heading into Selection Monday.

14. Ithaca (12-5-2) - Even assuming Rochester jumps Ithaca, the Bombers are likely safe too, with a nearly identical profile: 0.684 winning percentage (Rochester’s is 0.686), a 0.565 strength of schedule (compared to 0.571), and three wins and a draw in seven games versus ranked teams (compared to 4-3-0). Ithaca’s three ranked wins—along with a 0.500 record versus ranked—is probably enough. But that’s assuming Ithaca can hold off Clarkson in the rankings, despite the Bombers losing to Clarkson for a second time last week. The saving grace may be that Vassar’s head-to-head edge is likely to keep it above Clarkson, and Ithaca is likely to stay above Vassar (and therefore Clarkson).

15. Gettysburg (11-7-1) - It’s almost inconceivable that a team with seven losses should feel good without an automatic bid to the tournament, but Gettysburg played an incredible ten games against ranked teams. That paid off in two ways. First, the Bullets’ 0.627 strength of schedule will weigh in their favor against every other team on the board. Second, they showed well in those games, picking up four ranked wins (most of anyone remaining) and a tie. Add in last week’s 1-0 win over Haverford, which should see Gettysburg jump the Fords in the Mid-Atlantic, and we see Gettysburg as a pretty safe pick.

RIGHT SIDE OF THE BUBBLE (3)

16. Hope (14-5-1) - This is going out on a limb a bit, because we’re predicting that Hope will jump Wheaton (Ill.), North Central (Ill.), and Kalamazoo in the final Central Region rankings. But Wheaton missed the CCIW tournament and North Central fell in the semifinals in penalties. That opened the door for Hope, which beat Kalamazoo in the MIAA tournament, which should propel it above all three teams. The Flying Dutchmen will also benefit from a balanced profile, with one of the strongest remaining winning percentages and strengths of schedule of the realistic Pool C teams.

17. Ramapo (11-2-5) - The loss to Rowan will leave the Roadrunners nervous, given they have just the one ranked win. But the winning percentage is quite high compared to the remaining realistic Pool C teams, and the strength of schedule is just fine. While it will probably be a close thing, odds are Ramapo will be playing next weekend.

18. Kalamazoo (11-4-2) - Kalamazoo shouldn’t sweat too much, despite the loss to Hope, as it has a similarly balanced profile highlighted by three ranked wins and a 0.500 record versus ranked—something few teams remaining can claim.

SQUARELY ON THE BUBBLE (PICK 'EM 3 OF 6)

19. WPI (12-3-4) - The one ranked win is problematic, and it win came against Endicott way down at eleventh in the New England region. In other words, WPI really hasn’t beaten anyone of note. But the winning percentage is quite high for the remaining teams and the strength of schedule is strong enough. The New England region has valued WPI all year, and it might just be strong enough to pull five bids—including one for the Engineers.

20. Colorado College (14-6-0) - The Tigers are difficult to evaluate. They played (and won) four games against non-Division III opponents, meaning their DIII winning percentage (0.625), which is part of the primary criteria, is much lower than their actual winning percentage (0.700). That and a pedestrian strength of schedule could be the difference. But we think the Tigers have a good shot at a bid for two reasons. First, with 21 at-large bids, it seems inconceivable that any region will miss out entirely. The West has fallen victim to that before, but usually that’s because the West’s Pool C teams have no ranked wins or awful strengths of schedule. Colorado College’s strength of schedule is fine and it has two ranked wins, so it’s right there in the mix. Which raises the second point: Colorado College fares well on criteria beyond winning percentage, strength of schedule, and record versus ranked. For example, the Tigers beat Luther, while St. Thomas and Loras both lost to Luther. And if it comes to secondary criteria, Colorado College’s “out of conference” schedule—including games against Luther, Kenyon, and Ohio Wesleyan—is very strong.

21. Otterbein (13-7-0) - Otterbein’s win over Mount Union was probably enough for the Cardinals to jump the Purple Raiders. It evened the season series, and Otterbein has a decent edge in strength of schedule and an extra ranked win. The question will be whether the committee favors three ranked wins, even when matched against seven losses to ranked teams.

22. Roanoke (14-2-3) - The winning percentage is great, and a 0.500 record in four games versus ranked teams is solid, but the lone ranked win and the low strength of schedule probably dooms the Maroons.

23. Haverford (9-6-3) - The winning percentage is weak, but so is almost everyone else’s at this point. Haverford’s best shot is for the committee to greatly value strength of schedule, as the Fords have one of the strongest in all of Pool C (0.620). Two ranked wins and a decent showing in eight games versus ranked teams also provide evidence that Haverford can hang with strong teams in the tournament.

24. Vassar (9-7-2) - Clarkson would have a solid chance at making the tournament, except that it’s hard to see how they jump a Vassar team that won the head-to-head meeting and has a substantially better strength of schedule. Vassar’s two ranked wins and that strength of schedule also could it a puncher’s chance, but given that the Brewers only won half their games we think two probably isn’t enough.

WRONG SIDE OF THE BUBBLE

St. Thomas (12-4-3) - The winning percentage is solid for the bubble and the strength-of-schedule is excellent. The glaring hole in St. Thomas’s resume is its 0-4-2 record versus ranked teams. Compare that to North Central, for example, which has an extra blemish and a slightly weaker overall schedule, but went 2-2-0. The Tommies had six chances to pick up a ranked win and just couldn’t do it, probably dooming their season.

Loras (15-6-1) - This depends on how the North Region committee ranks its teams in the final rankings. Loras might have a chance, given the solid winning percentage of those teams remaining and a very good strength of schedule—even with just one ranked win. But the Duhawks are stuck behind St. Thomas, so they won’t even be up for discussion unless the Tommies make the field (unlikely) or they jump in the final rankings. No good reason for Loras to jump St. Thomas—both teams won a conference semifinal and lost in the final, and St. Thomas actually closed the strength-of-schedule gap to match Loras. But the regional committee could choose to rank Loras higher, thinking the Duhawks have a stronger at-large profile. If Loras makes the field, that’s a possible explanation.

Clarkson (11-5-4) - Similarly, Clarkson’s fate could depend on the East region committee. If that committee moves Clarkson above Vassar, then the Golden Knights have a fighting chance. If it doesn’t, as we suspect, Clarkson probably never even comes up for discussion.

Brandeis (10-5-4) / Trinity (Texas) (11-6-2) - The strength of schedule is solid for both teams, but the winning percentages are low and each squad has just one ranked win. Hard to see that being enough for a bid, particularly when teams like Vassar and Otterbein have similar but slightly better profiles.

Dickinson (10-8-0) - The strength of schedule (0.630) is the best overall in Pool C, and three ranked wins are often enough to warrant a bid. But the winning percentage is awful, and eight losses is too many, particularly when the Red Devils are unlikely to jump Haverford, another team firmly on the bubble.

Oglethorpe (14-4-0) / Wheaton (Ill.) (10-4-4) - Each team has its strength, Oglethorpe a high winning percentage and Wheaton a high strength of schedule. The problem is that each team is stuck behind several other bubble teams from its region—Oglethorpe behind Rowan, Ramapo, and Roanoke in the South Atlantic, and Wheaton (Ill.) behind Hope (we think), Kalamazoo, and North Central. That makes it likely neither team comes up for discussion. Even if they make it onto the table, each team has a big hole, Oglethorpe with its strength of schedule and Wheaton with its lone ranked win.

Mount Union (10-4-5) - A win over Otterbein probably would have seen the Purple Raiders make the field, as it would have given them a third ranked win, and strength-of-schedule increase from a conference final, and a head-to-head double over Otterbein. But the loss probably drops Mount behind Otterbein, and given Otterbein’s status, it’s hard to picture Mount getting a bid.

Comments or feedback for the authors?  Email D3soccer.com.