It's time to start up this thread. Why? The NCAA has just published its Soccer Championship Pre-Championship 2015-16 Manual; This past Monday, the NCAA published its initial RPI rankings and ratings; This week, the Regional Advisory Committees and the Women's Soccer Committee held the first of their every-other-week conference call meetings as part of their process of ultimately making at large selections and seeding teams for the bracket. Starting next week, I'll be publishing, with my weekly RPI report, a list of teams that, based on recent history, are: Within the range of potential #1 seeds; Within the range of potential #2 seeds; Within the range of potential #3 seeds; Within the range of potential #4 seeds; Virtually assured of getting at large selections if they are not conference champions; and Within the range of potential at large selections. For those who are interested in the criteria for at large selections, in the seeds, and in the process leading up to the bracket announcement, I suggest you go to these RPI for Division I Women's Soccer webpages: At Large Selections and Seeds, Criteria and/or Considerations Bracket Formation Process and Timeline As an additional resource, I have identified historic patterns that the at large selections and seeds have met, in relation to the criteria and teams' season results. Some simple examples are that, using the end-of-regular-season results, the #1 ranked team in the Adjusted RPI always has received a #1 seed, teams that are not conference champions but that are ranked #34 or better by the ARPI always have received at large selections, and teams ranked #58 or poorer never have received at large selections. There also are standards involving pairings of data such as conference average ARPI and team conference regular season and conference tournament finishing positions. These patterns provide a basis for identifying what decisions the Committee will make if it follows historic patterns as well as for identifying when the Committee goes off the beaten path. If you're interested in these patterns, go here for at large selections: At Large Selection Patterns and go here for seeds: Seed Patterns And, if you're interested, when looking at weekly RPI rankings, in the weekly ranges that teams have been in that ultimately have gotten at large selections and seeds, go here: Ranges to Apply to Weekly RPI Rankings Altogether, these resources should give you an idea of who's realistically in play for at large selections and seeds as we go through the season; who's a safe bet for an at large selection or seed when we get to the end of the season; and who's at play for the last at large selections and seeds, and what the likely Committee debate points are, when the Committee is making its final tough decisions.
For fun, if I disregard for now the 0.500 winning record requirement to qualify for an at large selection and also disregard automatic qualifiers, here are likely seeds and at large selections based on teams' records through the Thursday, September 24 games, based on the Committee's historic patterns. At this point, not enough teams fit the patterns to fill out an entire 64-team bracket. Within their groups -- #1 seeds, #2 seeds, #3 seeds, #4 seeds, and at large selections -- the teams are arranged in alphabetical order. A 5 represents an at large selection. An interesting note, at least to me: Virginia, at this point, does not have a single positive result against a Top 60 team -- not a win, and not a tie. 1 BYU 1 Clemson 1 FloridaState 1 PennState 2 ArizonaU 2 NotreDame 2 Rutgers 2 Stanford 3 Duke 3 NorthCarolinaU 3 SantaClara 3 WashingtonU 4 CaliforniaU 4 FloridaU 4 IllinoisU 4 SouthCarolinaU 5 Auburn 5 BostonCollege 5 ConnecticutU 5 Georgetown 5 Hofstra 5 IndianaU 5 KentuckyU 5 LongBeachState 5 LoyolaMarymount 5 LSU 5 MarylandU 5 MichiganState 5 MichiganU 5 MinnesotaU 5 MississippiU 5 MissouriU 5 NebraskaU 5 Northeastern 5 NorthTexas 5 NorthwesternU 5 OhioState 5 OklahomaU 5 Pepperdine 5 PortlandU 5 Purdue 5 SanDiegoState 5 Seattle 5 SouthAlabama 5 SouthernCalifornia 5 SouthFlorida 5 StJohns 5 Syracuse 5 TennesseeU 5 TexasA&M 5 TexasTech 5 Tulsa 5 UCF 5 UNCWilmington 5 VirginiaTech 5 VirginiaU 5 WashingtonState 5 WestVirginiaU
If you get to looking over this week's RPI rankings, here are the historical ranges (meaning, based on what's happened over the last eight years) for seeds and at large selections, given where we are in the course of the season: #1 seed range: teams ranked #22 or better #2 seed range: teams ranked #42 or better #3 seed range: teams ranked #85 or better #4 seed range: teams ranked #85 or better At risk of not getting an at large selection: teams ranked #11 or poorer Having the possibility of getting an at large selection: teams ranked #148 or better Thus, at large bubble: teams ranked #11 through #148
Following on my previous post -- disregarding for now the 0.500 winning record requirement to qualify for an at large selection and also disregarding automatic qualifiers -- here are likely seeds and at large selections based on teams' records through the Sunday, September 27 games, based on the Committee's historic patterns. At this point, not enough teams fit the patterns to fill out an entire 64-team bracket -- thus there are only 52 teams on the list. Within their groups -- #1 seeds, #2 seeds, #3 seeds, #4 seeds, and at large selections -- the teams are arranged in alphabetical order. A 5 represents an at large selection. At this point, this mostly is fun and to see how things change as the season progresses. I'm guessing there will be some very significant changes. 1 Clemson 1 FloridaState 1 MinnesotaU 1 PennState 2 BYU 2 NotreDame 2 SantaClara 2 Stanford 3 ArizonaU 3 NorthCarolinaU 3 Rutgers 3 SouthFlorida 4 NorthwesternU 4 SouthCarolinaU 4 VirginiaTech 4 WestVirginiaU 5 Auburn 5 BostonCollege 5 CaliforniaU 5 CalPoly 5 ColoradoU 5 ConnecticutU 5 Duke 5 FloridaU 5 Georgetown 5 Hofstra 5 IllinoisU 5 IndianaU 5 KansasU 5 KentuckyU 5 LongBeachState 5 LoyolaMarymount 5 LSU 5 MichiganState 5 MississippiU 5 MissouriU 5 NebraskaU 5 OhioState 5 OklahomaU 5 Pepperdine 5 SouthAlabama 5 SouthernCalifornia 5 StJohns 5 TennesseeU 5 TexasA&M 5 TexasTech 5 Tulsa 5 UCF 5 UNCWilmington 5 VirginiaU 5 WashingtonState 5 WashingtonU
Continuing along for this week -- disregarding for now the 0.500 winning record requirement to qualify for an at large selection and also disregarding automatic qualifiers -- here are likely seeds and at large selections based on teams' records through the Sunday, October 4 games, based on the Committee's historic patterns. At this point, not enough teams fit the patterns to fill out an entire 64-team bracket -- thus there are only 49 teams on the list. Within their groups -- #1 seeds, #2 seeds, #3 seeds, #4 seeds, and at large selections -- the teams are arranged in alphabetical order. A 5 represents an at large selection. At this point, this still mostly is fun and to see how things change as the season progresses. I'm guessing there will be some very significant changes. 1 FloridaState 1 MinnesotaU 1 PennState 1 Stanford 2 Auburn 2 Clemson 2 Rutgers 2 SouthFlorida 3 ArizonaU 3 NorthCarolinaU 3 VirginiaTech 3 WestVirginiaU 4 BYU 4 NorthwesternU 4 SouthCarolinaU 4 WashingtonU 5 BostonCollege 5 CaliforniaU 5 Cincinnati 5 Duke 5 FloridaU 5 Hofstra 5 IllinoisU 5 IndianaU 5 KentuckyU 5 LongBeachState 5 LoyolaMarymount 5 LSU 5 Marshall 5 MichiganState 5 MichiganU 5 MississippiU 5 MissouriU 5 NebraskaU 5 Northeastern 5 NotreDame 5 OhioState 5 OklahomaU 5 SantaClara 5 SouthAlabama 5 SouthernCalifornia 5 StJohns 5 TennesseeU 5 TexasA&M 5 TexasTech 5 UNCWilmington 5 VirginiaU 5 WashingtonState 5 WisconsinU
With another week's games in the system (through October 11) and still disregarding the 0.500 winning record requirement to qualify for an at large selection, here are likely seeds and at large selections based on the Committee's historic patterns. At this point, I come up with a list of 43 teams that look like they would be "in" if the Committee were deciding today. Fortuitously, that group includes teams from 10 conferences, so if I assume the top team from each conference on my list will be the conference's automatic qualifier, this leaves 21 automatic qualifiers that are not on my list of 43 teams. Adding my 43 and those 21 together gives a full bracket of 64 teams. In my list, within their groups -- #1 seeds, #2 seeds, #3 seeds, #4 seeds, and at large selections -- the teams are arranged in alphabetical order. A 5 represents an at large selection. This remains mostly fun and to see how things change as the season progresses. But although significant changes still are likely, things are starting to shape up a little bit. 1 Auburn 1 Clemson 1 FloridaState 1 PennState 2 FloridaU 2 NotreDame 2 VirginiaU 2 WestVirginiaU 3 MinnesotaU 3 NorthCarolinaU 3 SantaClara 3 SouthFlorida 4 BYU 4 SouthCarolinaU 4 Stanford 4 VirginiaTech 5 ArizonaU 5 BostonCollege 5 CaliforniaU 5 ConnecticutU 5 Duke 5 Hofstra 5 IllinoisU 5 KentuckyU 5 LongBeachState 5 LoyolaMarymount 5 LSU 5 MichiganU 5 MississippiU 5 MissouriU 5 NebraskaU 5 NorthwesternU 5 OhioState 5 Rutgers 5 SouthernCalifornia 5 StJohns 5 TennesseeU 5 TexasA&M 5 TexasTech 5 UCF 5 WashingtonState 5 WashingtonU 5 WisconsinU
Forgive me, this is a repeat of an entry I just put on the 2015 RPI thread, but I want to be sure it reaches those who are interested: For those interested, at the RPI for Division I Women's Soccer website, I've gotten rid of the old Blog page that didn't work very well and set up a new one that hopefully will work better. Regularly during the Fall soccer season, and from time to time during the rest of the year, I'll enter blog posts onto the page that I believe will be of interest to Division I Women's Soccer coaches and to fans with serious interest in the RPI and the NCAA Division I Women's Soccer Committee's NCAA Tournament decision-making process. I'll also answer questions that I believe will be interesting to coaches and serious fans. I've now set up the Blog page so anyone interested can subscribe to receive posts through a web-based news reader or via email. The instructions on how to do it are at the top of the page. If you're interested, here's a link to the page: RPI for Division I Women's Soccer Blog.
With another week's games in the system (through October 18) and still disregarding the 0.500 winning record requirement to qualify for an at large selection, here are likely seeds and at large selections based on the Committee's historic patterns. I've cut off the at large selections at the point necessary in order to allow room in the 64-team bracket for the automatic qualifiers that are not among the highly ranked teams. In my list, within their groups -- #1 seeds, #2 seeds, #3 seeds, #4 seeds, and at large selections -- the teams are arranged in alphabetical order. A 5 represents an at large selection. 1 Clemson 1 FloridaState 1 PennState 1 Rutgers 2 Auburn 2 Duke 2 FloridaU 2 VirginiaU 3 CaliforniaU 3 SouthFlorida 3 Stanford 3 WestVirginiaU 4 BYU 4 NotreDame 4 SantaClara 4 VirginiaTech 5 ArizonaU 5 ConnecticutU 5 Georgetown 5 Hofstra 5 IllinoisU 5 KentuckyU 5 LSU 5 MichiganU 5 MinnesotaU 5 MississippiU 5 MissouriU 5 NebraskaU 5 NorthCarolinaU 5 NorthwesternU 5 OhioState 5 SouthAlabama 5 SouthCarolinaU 5 SouthernCalifornia 5 StJohns 5 TexasA&M 5 TexasTech 5 UCF 5 UNCWilmington 5 WashingtonState 5 WashingtonU 5 William&Mary 5 WisconsinU
Sorry all, but my "likely" seeds and at large selections that I posted yesterday were off. I forgot a process step before telling my computer to spit out the numbers. Here is a corrected list. This time, I'm putting it in a sort of order - not for the seeds really although for them the order means something, but in order for the non-seeds. The list likely includes 10 conference champion automatic qualifiers. It's cut off at 43 teams, to leave room in the 64-team NCAA Tournament bracket for 21 additional automatic qualifiers. I suspect that the Colonial and Big West teams on the list will see their positions degrade over the next few weeks, as the teams from the power conferences get more opportunities to accumulate good results against highly ranked teams (i.e., the teams from their own conferences). As a note, in the outer edges of each seed pod and the at large selections, at this point the system I'm using requires me to make some judgments. They're fairly mechanical; and hopefully they'll be gone by the end of conference tournament play. 1 FloridaState 1 VirginiaU 1 Rutgers 1 PennState 2 FloridaU 2 WestVirginiaU 2 Clemson 2 Auburn 3 Duke 3 NotreDame 3 VirginiaTech 3 Stanford 4 SouthCarolinaU 4 SouthFlorida 4 NorthCarolinaU 4 ConnecticutU 5 MinnesotaU 5 MississippiU 5 WisconsinU 5 TexasA&M 5 CaliforniaU 5 BYU 5 ArizonaU 5 WashingtonU 5 SantaClara 5 MichiganU 5 MissouriU 5 NebraskaU 5 OhioState 5 IllinoisU 5 SouthernCalifornia 5 TexasTech 5 William&Mary 5 StJohns 5 NorthwesternU 5 CalStateFullerton 5 UCF 5 WashingtonState 5 LSU 5 Hofstra 5 KentuckyU 5 Cincinnati 5 UNCWilmington
OK, CP... Straighten me out here if I'm missing something critical. Given your statement that teams of ARPI or 34 or lower always make the tournament and 58 or above never make an at large selection (the bubble falls between), then based on the current ARPI the following conferences will have entries (my prediction is the first number): 7 ACC ........ 7 locks and one bubble 7 SEC ........ 5 locks and 3 bubbles 7 Big 1o ..... 4 locks and 4 bubbles 6 PAC-12.... 6 locks and no bubble 3 American Athletic ...3 locks and 1 bubble 3 Colonial .... 2 locks and 2 bubbles 3 Big 12 .... 2 locks and 2 bubbles 2 Big East... 2 locks and 1 bubble 2 West Coast... 1 lock and 1 bubble A total of 40 teams from the above 9 conferences. There are 23 remaining conferences whose champions will be in the NCAA for a total of 63. This is interesting for me to contemplate. I chose a total of 9 teams to add to the locks to get my personal prediction of how many teams the "multiple entry" conferences would have. By doing so, when the other 23 conferences were added in, that left only one position remaining to fill the bracket, for a total of 10 bubble selections. If it is indeed all fairly mechanical, and only ten slots are actually wrestled over, then I can certainly understand your fascination with trying to divine how it is that the committee makes the actual bubble decisions. Assuming that your historic "34 and below and 58 and above" remains constant, the committee is choosing ten bubble teams out of 20 candidates.
I'm not sure I completely understand your post, but I think you're about right. The only thing is, it looks like you're talking about where the teams are now. The 35 to 57 as the "bubble teams" (which may include some AQs, so it actually may be fewer teams) is as of the end of the season. So, any use of those numbers now (hopefully, I'll post my this week's results tomorrow) is more for educational purposes to see how things develop over the course of the season, than for representing what the Committee will do once the end of the season arrives. BUT, we are getting close to the end of the season, so we now are getting more direct views and fewer shadows. (I must be really tired to have come up with that description.) Oh, what the heck, here's what my system comes up with this week, for the 64 team field if decided on today with the Committee following precedent established over the last eight years and with the team currently at the top of each conference's rankings being the conference's automatic qualifier. (For the AQs, if two teams are tied at the top of a conference, the first in the alphabet is the one I give the AQ designations.) What you're seeing is seed (#1 through #4, not necessarily in order; #5 as at large selection, in order; and AQ = automatic qualifier). The order of information is seed or at large selection/AQ/ARPI rank/Team 1 AQ 2 VirginiaU 1 3 PennState 1 4 Rutgers 1 1 FloridaState 2 AQ 6 FloridaU 2 AQ 7 WestVirginiaU 2 10 NorthCarolinaU 2 AQ 5 Stanford 3 11 VirginiaTech 3 13 Auburn 3 8 Clemson 3 14 ArizonaU 4 AQ 9 ConnecticutU 4 19 CaliforniaU 4 15 MississippiU 4 18 SouthFlorida 5 17 NotreDame 5 12 Duke 5 25 MinnesotaU 5 29 TexasA&M 5 30 SouthCarolinaU 5 16 SantaClara 5 23 WashingtonU 5 28 BYU 5 49 MichiganU 5 22 UCF 5 41 NorthwesternU 5 37 OhioState 5 26 SouthernCalifornia 5 27 TexasTech 5 31 William&Mary 5 43 KentuckyU 5 42 NebraskaU 5 51 MissouriU 5 44 LSU 5 33 Georgetown 5 24 WashingtonState 5 39 UNCWilmington AQ 34 WisconsinU AQ 20 Hofstra AQ 32 CalStateFullerton AQ 35 StJohns AQ 38 Princeton AQ 21 SouthAlabama AQ 36 LoyolaMarymount AQ 46 NorthTexas AQ 54 SanDiegoState AQ 75 BostonU AQ 95 FloridaGulfCoast AQ 139 Milwaukee AQ 70 Seattle AQ 89 FairleighDickinson AQ 73 Samford AQ 79 BallState AQ 87 Liberty AQ 101 IdahoU AQ 177 Hartford AQ 218 Drake AQ 80 Monmouth AQ 115 MurrayState AQ 142 NorthDakotaState AQ 157 StephenFAustin AQ 258 JacksonState AQ 45 GeorgeWashington
Very interesting indeed. Thanks. Yes I was very clear that my post only involved present stats and that there will be changes in 10 days that the committee will be working with, However, I was using the present stats as representative to make the conclusion that the committee will then only be looking at the non AQs in the 35 through 57 group to choose the remaining at large selections and the number is apparently not that large. My math may have been lacking, but I was interested to see what kind of process they would go through when their choices are limited by the historic cut-offs you have noticed over the years. But never mind my rambling, thanks for your post! (I'll wait until the shouting is over in a couple of weeks.) Cheers!
I think I get it. You're right that, from history, the teams from 35 to 57 are the teams that are (1) at risk of not getting an at large selection but with the possibility of getting one. And, if you take out the automatic qualifiers that are in that range, then you end up with the Committee really looking at a fairly small group of teams, maybe 15 to 20.
Here's what my numbers say the likely seeding and at large selections would be following the Sunday, November 1 games. I've also indicated the Automatic Qualifiers, based on teams' regular season conference standings, so this gives a picture of a potential Tournament field. What you're seeing is seed or at large selection (5)/AQs/ARPI Rank/Team. The at large selections are in order starting with the most clearly in: 1 AQ 5 VirginiaU 1 AQ 2 PennState 1 3 Rutgers 1 1 FloridaState 2 AQ 8 FloridaU 2 AQ 6 WestVirginiaU 2 AQ 4 Stanford 2 11 NorthCarolinaU 3 10 NotreDame 3 9 Clemson 3 12 Auburn 3 15 SouthFlorida 4 AQ 7 ConnecticutU 4 13 VirginiaTech 4 17 ArizonaU 4 19 BYU 5 14 Duke 5 20 SantaClara 5 21 CaliforniaU 5 36 MinnesotaU 5 23 MississippiU 5 25 SouthCarolinaU 5 22 SouthernCalifornia 5 34 BostonCollege 5 28 TexasA&M 5 27 OhioState 5 46 MichiganU 5 24 UCF 5 41 NorthwesternU 5 40 WashingtonU 5 39 LSU 5 54 IllinoisU 5 29 William&Mary 5 33 TexasTech 5 47 WisconsinU 5 32 Georgetown 5 50 MissouriU 5 30 WashingtonState 5 42 OklahomaU
Oops, sorry, I forgot top include the unseeded Automatic Qualifiers. Here's the list again, with the unseeded AQs at the bottom. They include Hofstra and South Alabama. 1 AQ 5 VirginiaU 1 AQ 2 PennState 1 3 Rutgers 1 1 FloridaState 2 AQ 8 FloridaU 2 AQ 6 WestVirginiaU 2 AQ 4 Stanford 2 11 NorthCarolinaU 3 10 NotreDame 3 9 Clemson 3 12 Auburn 3 15 SouthFlorida 4 AQ 7 ConnecticutU 4 13 VirginiaTech 4 17 ArizonaU 4 19 BYU 5 14 Duke 5 20 SantaClara 5 21 CaliforniaU 5 36 MinnesotaU 5 23 MississippiU 5 25 SouthCarolinaU 5 22 SouthernCalifornia 5 34 BostonCollege 5 28 TexasA&M 5 27 OhioState 5 46 MichiganU 5 24 UCF 5 41 NorthwesternU 5 40 WashingtonU 5 39 LSU 5 54 IllinoisU 5 29 William&Mary 5 33 TexasTech 5 47 WisconsinU 5 32 Georgetown 5 50 MissouriU 5 30 WashingtonState 5 42 OklahomaU AQ 26 StJohns AQ 35 CalStateFullerton AQ 31 Princeton AQ 16 Hofstra AQ 18 SouthAlabama AQ 45 GeorgeWashington AQ 62 SanFrancisco AQ 52 SanDiegoState AQ 69 BostonU AQ 75 BallState AQ 91 FloridaGulfCoast AQ 70 Samford AQ 85 FairleighDickinson AQ 79 Liberty AQ 90 IdahoU AQ 134 Milwaukee AQ 154 StephenFAustin AQ 155 Hartford AQ 81 Monmouth AQ 109 MurrayState AQ 158 NorthDakotaState AQ 218 Drake AQ 248 MississippiValley AQ 48 NorthTexas AQ 60 Seattle
7 out of 33 at large bids from the Big Ten? I don't care what the numbers say. No way the BIG is that good. 8 from the ACC - fine. I've seen a ton of Big Ten soccer this year and the top five in the standings can compete with anyone on most days. OSU, Minnesota, and Illinois - not so much.
There still are more games to play, so maybe you're right. We'll see. One thing to remember, though, is that it doesn't matter what anyone has seen, in terms of teams actually playing. All that matters is their records, looked at in the various ways spelled out in the selection criteria. For the at large selections, the Committee is obligated to use the selection criteria and only those criteria.
Looking back, the Big Ten had a very good nonconference season overall. I could see them getting anywhere from 7-9!! If Nebraska can sneak in as well. Like cpthomas said, lots of soccer this week. Looking forward to the SEC Tournament today.
The Committee can go as far down the list as it wants. But, based on the last eight years, the poorest RPI rank to get a #1 seed has been 7.
There's an interesting situation that could occur and really put the Women's Soccer Committee to the test. South Alabama currently is ranked #18 in the ARPI. What if they don't win their conference tournament? Will they still get an at large selection? Not winning their conference tournament would mean they have suffered a loss or a tie after which they lost the shootout. This would drop them down on the RPI list, but not that much. I believe they'll still be well within the RPI ranking range that always has gotten at large selections. But look at rankings of the teams they've beaten, lost to, and tied: Won against: 61, 74, 77, 95, 102, 120, 123, 141, 208, 229, 233, 244, 247, 252, 274 Lost to: 13, 15 Tied: 80 So, what does their record demonstrate? That they can win against teams ranked #61 or poorer. When the Committee gets to applying the at large selection criteria, of which the RPI is only one, they're going to have to make a very tough decision. Except for its RPI rank, South Alabama does not have the credentials to get an at large selection. But, if the Committee declines to give them an at large selection, it will raise, in public, a major question about the legitimacy of the RPI. It's worth noting that my Iteration 5 URPI has South Alabama ranked #43 and my Elo system has it ranked #75. Iteration 5 corrects the RPI's problem with how it measures strength of schedule but, because it still has the basic RPI structure, it still tends to underrate teams from stronger conferences and to overrate teams from weaker conferences. Looking at who South Alabama played, there's a reasonable argument that the mid-point between the Iteration 5 rank and the Elo system rank is right about where South Alabama should be. That would put them at #59.
Yah, RPI also is sequence-insensitive: it ignores which subset of your foes you lost to. Elo is the opposite: it is highly sequence-sensitive. We need a Great Compromise, or something