Ohai has a sister that started on the USC team that won the NCAA in 2007, have there ever been two sisters win NCAA soccer titles from two different schools?
I'm guessing not, but I know that Yael Averbuch's sister Shira played as a little-used sub for the 2009-10 finals teams for Stanford. Last year should have been her fourth year of eligibility (and thus should would have been a national champion), but she wasn't on the roster.
It looks like Christine Nairn, Crystal Dunn and Caroline Miller are the three Hermann Trophy finalists. Congrats ladies!
Votes are already in but hopefully Dunn comes out on top. Hopefully the coaches didn't cast 1st place votes for a plyer that scores a bunch of goals in a season and each day knows her one position on the field - that's easy. Instead they should cast 1st place votes for the player that was the most complete player -- dominated on defense (another ACC defensive player of the year award which no one has ever repeated before her), dominated in the midfield when moved into that postiion for the first time for the NCAA tournament (5 goals in 4 games and hughe goals), key to the UNC national champtionship and (although not NCAA college stuff) key to the U20 team's world cup championship. That's a special player having a special year.
Give it to Crystal Dunn. She deserves it. Nice season for Nairn and Miller. But even better season for Dunn. Crystal always gets the job Done, for Crystal Dunn.
Im stunned Miller is on the list. She doesnt seem to garner much respect outside the ACC. In the TDS conference ranking Miller (#6) isnt even ranked above Morgan Brian (#4). Love to see Dunn get it. She did exactly what a star is supposed to do.
Top Drawer is the worst. They do very little research for the amount of info they crank out. If they want to cover women's soccer, have at it, just actually DO the work. Generating content and lists is a poor way to operate. They are especially bad for the lower ranked conferences, because again, it would require more legwork. Rant over.
Ask ACC defenders who they r happy to see graduate and I bet u hear Miller and McCarty but they dont seem to get as much respect out of the conference.
Belatedly, I've competed a full review of how the Women's Soccer Committee made its at large selections and seeds for the 2012 NCAA Tournament. It's given me some new insights into which teams find their way into the at large selection "bubble" and into the role that conference average strength and within-conference placement may play in the seeding process. It's complete with an Excel workbook that has lots of information about the relevant teams and my complete system for at large selections and seeds, developed to try to replicate the Committee's decision-making process. My system produces at large selections that match the Committee's and seeds that come very close to matching the Committee's and that let me exactly match the Committee's with the help of my new sense of the role of conference strength and within-conference finish in the seeding process. Here's a link to the "NCC Tournament: How Did the Committee form the 2012 Bracket" page at the RPI website. The Excel workbook is an attachment at the bottom of the page. I'll post a few of my thoughts on this here, shortly. https://sites.google.com/site/rpifo...t-how-did-the-committee-form-the-2012-bracket
SELECTION OF AT LARGE TEAMS At the webpage for which I provided a link in the preceding post, I explain in detail what I believe is the essence of the process the Women's Soccer Committee uses to identify the 34 teams (that are not conference champions) that will get at large slots in the NCAA Tournament. The Committee may not go through the process exactly the way I describe it, but I believe my process captures the substance of what the Committee does. In this post, I'll report on what teams my process leads to for at large selections. The process includes one revision from prior years, based on the Committee's not awarding Dartmouth an at large position. With that revision, my process leads to the same selections that the Committee made. Identification of the Bubble Group. Ordinarily, applying my process for identifying the bubble group, the group would have included the 15 teams from Arizona State (#39 ARPI) to William & Mary (#58 ARPI) that were not conference champions. (The conference champions within that range were Princeton, Cal State Northridge, Florida Gulf Coast, North Texas, and Utah State.) The Committee, however, did not give #37 Dartmouth an at large spot so it too must have been included in the bubble. (Miami OH at #38 was a conference champion.) The question is, Why did the Committee add Dartmouth to the bubble? To answer this question, I took a look at Dartmouth's actual results. Here are its results against teams in the top 58 (in other words, within the bubble group or better): As this table shows, Dartmouth's best result was a win against #52 South Florida, itself a bubble team. Further, Dartmouth had lost all its games against higher ranked teams, including #41 Princeton and #47 Rutgers. This led me to the conclusion that, as a supplement to my past method for identifying 15 bubble teams, I needed an additional step. This step is to look at the actual game results of all teams (that are not conference champions) on the "better" ranked side of the bubble. What I look for there is results that demonstrate that the team is able to compete at the rank level of the other teams on this side of the bubble. If a team's results demonstrate that it is able to compete at this level, then it is automatically "in" without going through the bubble evaluation. On the other hand, if a team's results don't demonstrate it can compete at this level, then it becomes an additional bubble team. Measured against the test I just described, Dartmouth did not pass the test. Thus I now add Dartmouth to the bubble group, bringing the group to 16 teams. From this group, I now must select 9 teams for at large positions. (Ordinarily, I would need to select 8 teams from a 15-team bubble, but the addition of Dartmouth brings the numbers to 9 and 16.) As a matter of interest, I also did a check of the actual results of all the other teams on the highly ranked side of the bubble to be sure they all passed the test, and they did with one exception. Stephen F. Austin, at #28, had no wins or ties against teams in the top 58. In this particular case, Stephen F. Austin was a conference champion, so it could not be treated as a bubble team. If it had not been a conference champion, however, it would have provided an interesting case for the Committee, since its #28 ARPI rank is much better than any rank the Committed has denied an at large position in recent years. Nevertheless, if it had not been a conference champion then using my system it would have become a bubble team and would not have received an at large position! My best guess is that the Committee would have reached the same conclusion and not given it an at large position. One might wonder how Stephen F. Austin could have an ARPI of #28, yet no wins or ties over teams in the top 58. The following table, showing Stephen F. Austin's entire season record, gives a clue as to how this can happen: This table shows that, in fact, Stephen F. Austin's best result was a win against TCU, ranked #66 by the ARPI as indicated in the "Opponent's Rank" column. The "Opponent's Contribution to Strength of Schedule Rank" column, however, shows the rank of each opponent in terms of the value of its contribution to Stephen F. Austin's strength of schedule. (SoS has an effective weight of 50% in determining a team's RPI.) As the table shows, there can be significant differences between teams' ARPI ranks and their Contribution to Strength of Schedule ranks. (Kolabear has pointed this out many times.) The table also shows that for Stephen F. Austin, these differences caused its opponents' strength of schedule contribution to its ARPI to be significantly higher than the opponents' ARPIs would justify: as an entire group the opponents' strength of schedule contribution ranks were 213 rank positions higher than their ARPIs would justify, and on average each opponent's strength of schedule contribution rank was 10.65 rank positions higher than justified. Thus although on the surface it might have looked radical to deny Stephen F. Austin an at large position given its #28 ARPI rank, a deeper look suggests that Stephen F. Austin was substantially overrated by the ARPI due to the RPI's significant overstatement of its strength of schedule. Making the At Large Selections At the RPI for Division I Women's Soccer website, on the page with the link in the preceding post, I describe in great detail the system I use for making at large selections from the bubble group. In summary, it focuses on head to head results, results against common opponents, and results against teams within, or on the "better ranked" side of, the bubble. The only place that calls for subjective judgment is in ranking the teams based on their results against teams within or on the better ranked side of the bubble. The system gives each team a Total Score using a "round robin" type of comparison. Here is how my system assigned Total Scores to the teams: Washington State 12 Illinois 12 Long Beach State 11 Miami FL 11 Auburn 9 Wisconsin 8 Colorado College 8 Arizona State 7 Rutgers 7 ------------------------------ Oregon State 6 Louisville 5 South Florida 3 Minnesota 1 Dayton 1 Dartmouth 0 William & Mary 0 Thus in my system, the top 9 teams (through Rutgers) get at large selections. The teams from Oregon State onward do not. This exactly matches the Committee's decision. At the time of the Committee's at large selections, there was some discussion about why Arizona State received an at large selection rather than Oregon State. The case of those two teams provides a good way to illustrate how the process works. The "round robin" process involves a one v one comparison of each bubble team to each other bubble team. Thus one of the comparisons was Arizona State to Oregon State. As the numbers from the full round robin worked out, whichever of these two teams won the comparison with the other was the one that was going to receive the at large selection. Arizona State won the comparison, giving it its 7th point, leaving Oregon State with 6. If Oregon State had won the comparison, then it would have had 7 points and been "in" and Arizona State would have had 6 and been "out." Here is why Arizona State won this critical "point": First, regarding results against the bubble teams and the other teams on the better side of the bubble, here are the two teams' results: Arizona State results: Oregon State results: Comparing these results, I had Oregon State ranked ahead of Arizona State in their competition with each other, giving it one "Selection Point" in that competition. Second, looking at head to head results between Oregon State and Arizona State, the two teams tied at Oregon State. In my system, I treat an away tie as equivalent to a loss by the home team and a win by the away team. I decided to treat tie games that way after thinking about it quite a bit and taking into consideration work I've done on the extent of home field advantage. Specifically, I assume a team with a tie at home would have lost at a neutral site; and a team with a tie away would have won at a neutral site. And, I assume a team with a tie at a neutral site would have won at home and lost away. I use this same approach when I get to common opponent results. Others might disagree with this approach to ties, but it is how I do it and it produces results that match the Committee's. Thus Arizona State gets one head to head Selection Point in the competition with Oregon State. Third, looking at results against common opponents of the two teams (where two teams did not have the same result): Against Washington State, Arizona State won away and Oregon State lost away. Selection Point to Arizona State. Against Washington, Arizona State lost away and Oregon State tied away. Selection Point to Oregon State. Against Southern California, Arizona State won at home and Oregon State tied at home. Selection Point to Arizona State. Against Oregon, Arizona State won away and Oregon State lost away. Selection Point to Arizona State. Thus for common opponents, Arizona State has a net of 2 Selection Points. Altogether, this gave Arizona State 3 Selection Points compared to Oregon State's 1, giving Arizona State a round robin point in its comparison to Oregon State and bringing its "Total Score" to 7 as compared to Oregon State's "Total Score" of 6.
#1 SEED SELECTIONS For the #1 seed selections, the candidate pool is the top 6 ARPI teams. Running those teams through the process produced the following "Total Point" results: Stanford 4Florida State 3BYU 2Penn State 1--------------------------San Diego State 1UCLA 0 With Penn State and San Diego State each receiving one "Total Point," this left the question of which should receive the fourth #1 seed. They had no head to head or common opponent results. Here are their results against bubble or better than bubble teams: San Diego State results: Penn State results: Comparing these results, I ranked Penn State's results as better than San Diego State's. I'll emphasize that in my ranking of results, I am mostly concerned about good wins and ties and am less concerned with losses. My ranking Penn State higher means that I considered Penn State's record to have showed it was capable of competing at a higher level (win over #7 Virginia and #12 Ohio State, for example) than San Diego State (effective win over #11 Santa Clara and win over #19 Portland). Thus Penn State gets the fourth #1 seed, based strictly on results against bubble or better teams. These four #1 seeds match the Committee's decision.
#2 SEED SELECTIONS The candidate pool for the four #2 seeds, in my system, is the two teams from the #1 candidate pool that did not get #1 seeds plus the #7 through #13 ARPI teams. Running these teams through my system produced the following "Total Point" results: Virginia 7Florida 6San Diego State 5North Carolina 5-------------------------Santa Clara 3UCLA 3Ohio State 1Marquette 1Baylor 1 In my system, the four #2 seeds go to the top four on this list. This matches the Committee's decision.
#3 SEED SELECTIONS The candidate pool for the four #3 seeds, in my system, is the teams from the #2 seed candidate pool that did not get #2 seeds plus the #14 through #18 ARPI teams. Running these teams through my system produced the following "Total Point" results: Duke 7Wake Forest 7UCLA 6Baylor 4------------------------Ohio State 4Tennessee 4Marquette 2Santa Clara 2Texas A&M 1UCF 0 This left the question of whether to give Baylor, Ohio State, or Tennessee the fourth #3 seed. In terms of head to head and common opponent results, only Baylor had a 1 point advantage over Tennessee in common opponent results. In my rankings of results against bubble or better teams, I had Ohio State ranked #4, Baylor #5, and Tennessee #6, although the three of them were very close. Altogether, this created some basis for giving Baylor the fourth #3 seed, although I also could see giving it to Ohio State or Tennessee. The Committee actually gave the fourth #1 seed to Baylor. I had thought, in 2011, that the Committee might have taken conference strength into consideration in some of its seeding, so I decided to look at conference strength, conference placement, and the conferences of the teams that already had been seeded. The following list has the top 8 conferences in terms of average ARPI, in order, together with the number of seeds each conference already had received (including the top 3 from the #3 seed candidate list above): 1. ACC -- 52. Pac 12 -- 23. Big 10 -- 14. SEC -- 15. Big 12 -- 06. West Coast -- 17. Big East -- 08. Mountain West -- 1 In addition, I looked at where the three teams -- Baylor, Ohio State, and Tennessee -- finished within their conferences. (I explain how I calculate conference finishing position on the linked webpage several posts up. Essentially, I average conference regular season finish position and conference tournament finish position.) Baylor 1.5Ohio State 1.5Tennessee 4.75 Looking at this conference-related information as a whole, what stands out is that so far, the Big 12 has no seeds. With other things being relatively equal, I give the fourth #3 seed to Baylor since it's from the Big 12. With Baylor getting the fourth #3 seed, my system's seed decisions match the Committee's.
#4 SEED SELECTIONS The candidate pool for the four #4 seeds, in my system, is the teams from the #3 seed candidate pool that did not get #3 seeds plus the #19 through #25 ARPI teams. Running these teams through my system produced the following "Total Point" results: Maryland 10 Portland 9 Pepperdine 8 Tennessee 7 Boston College 7 Ohio State 6 Marquette 5 Santa Clara 4 Texas A&M 3 Notre Dame 3 Texas Tech 2 Michigan 2 UCF 1 The Committee gave #4 seeds to Maryland and Portland, which matched my system. On the other hand, my system -- without going through a conference-based analysis such as I did for the last #3 seed -- would have given a #4 seed to Pepperdine and the other #4 seed to either Tennessee or Boston College, whereas the Committee gave those two #4 seeds to Ohio State and Marquette. So, I again look at conferences to see if there is an explanation there for the Committee's decision. If I give #4 seeds to Maryland and Portland leaving the last two #4 seeds open pending a conference analysis, here is how the numbers of seeds looked, by conference, with the conferences again in order of average ARPI: 1. ACC -- 6 2. Pac 12 -- 2 3. Big 10 -- 1 4. SEC -- 1 5. Big 12 -- 1 6. West Coast -- 2 7. Big East -- 0 8. Mountain West -- 1 9. Conference USA -- 0 In addition, here is where the potential #4 seeds ranked within their conferences (leaving out Notre Dame, Texas A&M, Texas Tech, Michigan, and UCF as having scored too poorly in the "Total Points" competition): Marquette 1 Ohio State 1.5 Pepperdine 3.5 Santa Clara 3.5 Tennessee 4.75 Boston College 7.5 Looking at the numbers of seeds the conferences already had and where the candidate teams finished within their conferences, if I am going to include a conference-based factor in my seeding decisions, then I give one of the #4 seeds to Marquette because the Big East does not yet have a seed and Marquette had the best conference finish of any of the teams, and I give the other #4 seed to Ohio State because the Big 10 is the third strongest conference but has only one seeded team so far and Ohio State had the next best conference finish of any of the teams. (I'll also note that Maryland's conference finish at 2.75 and Portland's at 2 also were better than the other candidates' finishes.) Thus when using a conference-based factor for the final seeds my system's seeds match the Committee's. What the final two #4 seeds and the last #3 seed suggest is that the Committee, before finalizing the #3 and #4 seeds, first checks conference strength, the number of the conferences' seeds among the teams already seeded, and conference finishing positions, to make sure that the seeds are distributed among the conferences fairly and rationally.