With a very little effort, you could have done this yourself, as the resources are at your fingertips. Here are the non-conference opponents of Virginia, with their RPI ranks. While the RPI has flaws, they are not so big as to change the overall tenor of the Virginia non-conference schedule. Virginia won all these games except for Penn State: Richmond 221 West Virginia 44 George Mason 223 George Washington 275 Santa Clara 36 James Madison 95 Penn State 19 (loss) Oklahoma 101 Compare this to North Carolina, which is one of its competitors for a #1 seed and which won all of its games: Washington 82 Arkansas 43 Illinois 115 Ohio State 38 Northwestern 103 Stanford 32 Florida 116 Overall, Virginia played a much weaker non-conference schedule.
I agree with this assessment. It's just a fact that Virginia played a weaker schedule in relation to other one seed contenders. However, I tend to agree with Chris Henderson that the committee will reward Virginia for winning the ACC regular season outright (and certainly the tournament too if they top FSU) and will likely end up punishing UNC for not making the ACC tournament. He has UVA as a #1 seed lock now.
The Duke game wasn't a good indication already? I think it's safe to say Wake was impressive again tonight. That was a fun game to watch. Hope their starting goal keeper recovers for the NCAAs. That was a nasty hit. Wake seems to be hitting their stride at the right time. Also I'm not sure their Louisville loss is considered "bad". Louisville, despite having a down year, still has an RPI hovering around bubble territory (though they certainly won't get in).
You dont know what the rankings of the teams will be when you schedule the games and you address one part of it which may or may not be overweight. Head to head is seemingly irrelevant. UVA beat Duke. UVA beat teams UNC did not (Duke and NC State) and yet non common opponents have more weight?
I'm basically agreeing with you. It's a fact that UVA's schedule ended up being a little weaker than the other teams in contention (not necessarily by design, though others seem to think this is a pattern. I'm not saying that.). Santa Clara ended up being a dud. West Virginia also a dud. Penn State, dud(ish). All of those were teams in the hunt for the college cup when the season began (and certainly when UVA scheduled them). What seems to be bringing UVA down is the teams with the 200+ RPI rank where they could have scheduled 100+ teams instead. I think UVA will still get a #1 seed.
From 2007 through 2019, North Carolina's average Non-Conference RPI Rank was 11. Virginia's was 16. This year currently, Duke is 1, Florida State 3, North Carolina 5, Rutgers (included because a possible or even likely #1 seed) 16, and Virginia 29. Virginia's 29 this year is very good and its 16 average prior to this year is excellent. However, we are discussing what it takes to get a #1 seed. Over the years since 2007, no team with an NCRPI rank of 29 or poorer has gotten a #1 seed -- although Virginia, even if it does not win against Florida State, may end up better than #29. With an average NCRPI rank of 16, Virginia clearly has high level non-conference schedules. The question is whether its non-conference scheduling meets what the Committee expects of a #1 seed. This year, its NC schedule is borderline. I am not arguing about whether Virginia should or should not get a #1 seed, as I try my best not to form opinions about what the Committee should or should not do. I simply am saying that based on past Committee decisions, it is unlikely they will get a #1 seed if they do not win against Florida State. I have not run a Virginia win in that game through my program to see if a win there would be enough to balance out their NC rank. It might or it might not. And, the Committee does not have to follow past precedent, especially if it sees a profile it has not seen before.
If UVA wins the ACC tournament or, tbh, even loses Sunday and UNC gets a 1 seed above them, the whole system should be blown up. It’s the underlying problem with numbers (I was a math major by the way). I wouldn’t make sense when you look at their resume. UVA will have finished with 1 or 2 losses, won the regular season title of the best conference in America, and also won the tournament of that same league (or lost in the final). They will have beaten Duke (which UNC couldn’t do), won at NC State (which UNC couldn’t do), won at Wake Forest, beat ND, beat Clemson, drew at UNC and at FSU and potentially have a win against FSU. They potentially lose out on a 1 seed all because a computer doesn’t like that they played Richmond on August 19th (I know it is more than that). I’m not arguing for the eye test or to stop looking at non-conference schedules. I would be arguing for more common sense use of numbers. In any discussion without the RPI of any sort, no one would say it makes sense for a 6th place team (UNC tied WF at 6) in a conference to be seeded over the top team in that same league. When comparing teams within the same conference, there has to be some consequence for where you finish in your league (unbalanced or not). Since people can’t interpret tone on the internet, this is not impugning Cpthomas’s work. They are only putting the data into formulas and giving us case history based on past results. It is impressive the amount of time they put in to do it. I would be impugning the method by which teams are seeded at the committee level. I agree with Chris Henderson, though. I think they will be a 1 in which case my rant was unnecessary, but it was enjoyable. Love these discussions.
when beating teams ranked 100 or 200 becomes a bigger factor than common results vs teams ranked much higher, then it feels like too much weight is being given to it.
Or it feels like they are putting guaranteed wins on their schedule. There is nothing inherently wrong with that until the competition is for a #1 seed compared to teams that played a tougher non conference schedule. That stat is used to promote linkage across legues regions, which UVA has chosen to avoid. NCRPI it the traditional measure the ACC uses to dominate the selection weekend. They played only one top 25 team out of conference and lost that match. In conference their records as 1 win, two ties against the top25, which is ok but not 1 seed material. a win against FSU could change that.
http://wosoindependent.com/college/2021/information or https://sites.google.com/site/rpifordivisioniwomenssoccer/Home pick your Chris
One note worth mentioning: For at large selections, the Committee is required to use the NCAA-specified factors. For seeds, although the Committee certainly considers them, they are not binding. Thus the Committee has more flexibility when seeding.
Sorry, UVA has two wins against top 25 in conference. I’m still wrapping my head around Notre Dame being an ACC team after all these years.
Not true. What cost uva is Santa Clara not doing well in non conference play. RPI 100 teams are no more likely to beat uva than RPI 200 teams. It’s window dressing.
im sure UVA lost ground with SCU not doing as well this year. That why you should schedule more top 25 projected teams. Copmplaints of window dressing don’t impress. in 2008 Santa Clara scheduled a super non conference season that included most of the top 10 teams in the country, then were brutally decimated by a series of injuries and won only 3 games. By the end of the season they could barely put a team on the pitch. Their strength of schedule was such that they still ended the season with an RPI of 96.
I generally agree with Number007’s thoughts about Virginia. And, head-to-head results and results against common opponents are factors the Committee considers. The reasons I come up with the possibility the Committee will not give Virginia a #1 seed are as follows. This does not mean I think it is what the Committee will (or should) do. It rather is a description of what past precedent suggests the Committee will do. 1. I evaluate how teams that have gotten #1 seeds have scored for each of a whole bunch of individual and paired factors (which include RPI rank, NCRPI rank, head-to-head results, results against common opponents, good results (wins and ties) against Top 50 opponents, and so on. For each factor, I look to see if there is a score where teams that had that score or better always have gotten #1 seeds. For each factor, I also look to see if there is a score where teams that had that score or poorer never have gotten #1 seeds. I do this for each of the Top 7 teams, because no team ranked 8 or poorer has gotten a #1 seed since 2007 when I began collecting data. 2. Once I have gone through that process, I end up with the following: a. Some teams have one or more YES scores (i.e., a score for which teams that have it always have gotten #1 seeds) and no NO scores. I identify these as clear #1 seeds (meaning that they clearly will get #1 seeds if the Committee follows its historic patterns); b. Some teams have one or more NO scores and no YES scores. I identify these as clearly not #1 seeds; c. Some teams have no YES and no NO scores. If there are fewer than 4 clear #1 seeds, I identify these as potential #1 seeds; and d. Some teams have one or more YES and one or more NO scores. These are teams that have profiles the Committee has not seen historically. I identify these also as potential #1 seeds. 3. If there are fewer than 4 clear #1 seeds, the next question is what is the best way to guess who the Committee will pick to fill the remaining #1 seed spots, from among the potential #1 seeds identified in items 2.c and 2.d above. To answer this question, I have done a study seeing which of all the individual and paired factors, if applied retroactively to all the years since 2007, would come closest to matching the Committee decisions of which teams to pick to fill #1 seed positions not clearly filled by the YES and NO factors. 4. For the #1 seeds, as it turns out, the Non-Conference RPI comes the closest of all the factors to filling those remaining #1 seed spots. In fact, as shocking as it may seem, when I retrospectively apply this process to the data from each year since 2007 (excluding 2020, for which I did not try it), it correctly selects as #1 seeds all but one of the teams that actually got #1 seeds. This does not mean it is what will happen this year. It simply says that based on a systematic study of the Committee’s decisions since 2007, the process is a best guess as to how the Committee as a whole (as distinguished from its individual members) thinks about #1 seeds and what that thinking would produce if applied to this year’s data. Looking at Virginia this year, as of games through November 5 and if it loses to Florida State, using the above system it is not a clear #1 seed (which is an indication of how rigorous the requirements are for a YES #1 seed score). This is what causes an evaluation of its Non-Conference RPI, where it does not score well in relation to the other potential #1 seeds. Thus the process I have described says the Committee, if it follows its historic patterns, will not give it a #1 seed. Frankly, however, I would not be at all surprised if Virginia becomes the second team since 2007 that the Non-Conference RPI does not work for as the deciding factor. If that is what happens and it gets a #1 seed, then that will get added to the data base and I will get to go back and see whether the Non-Conference RPI factor still is the best at matching the Committee’s past decisions or whether there now is a different factor that is better.
I very much agree with this statement! I don't do the whole "wonky analytics" (no disrespect to those who do), but I use a team's record. I look at how many top 25 or even top 10 teams they've beaten or lost to. I do the eye test. Which 2 or 3 teams would I least like my team to have to play? I promise you that UVA is a top 4 if not the #1 team in the USA. That is why I find it remarkable that some systems think there are 4 teams out there clearly better than them. What 4 teams are clearly better than UVA?
[ UVa played a very challenging schedule. Looking at a teams record now to determine how good they were earlier in the season makes little sense. I i told you team A would go 1-1-2 in its last four games vs ranked opponents. Conceding a goal in every game (5) and beating the lowest ranked team 2-1 in ot for the only win. Would you call them an unequivocal #1 over Team B who has gone 2-0-2 vs its last four and conceded 2g total? Current form has to be relevant in these things.
I took a quick look and there is a "secondary criteria" below for late-season trends but supposed to only be used when RPI, common opponents, and h2h don't yield a selection when comparing teams. There's also a secondary criteria of "results vs RPI 1-75" so your so-called quality wins vs bad losses. ● Late-season performance — defined as the last eight games including conference tournaments (strength and results). What you don't know is how much weight each committee member gives to these different criteria. The manual even states the criteria are 'not necessarily in priority order". So, I think the analysis of @cpthomas is awesome but some would argue just waiting for the announcement is fine too Also always true there is a ying/yang to this process. For every team that's trended pretty well and want that criteria to be used, there are those who don't and prefer the "body of work" argument.
Arguing over whether a P5 team like UVA, not the ACC title winner, is a 1 or a 2 seed is the most boring story line for me and I assume the rest of America. It happens to create the biggest controversy on this thread every year but I don't see the satisfaction in arguing it either way. We already know the committee is favoring the ACC over any other league whether or not they get the designation of top four. And I think the weight of a 1 vs 2 seed is way less impactful than the match-ups between here and the College Cup. For the rest of us here, can we talk about first round match-ups and what remains of the bubble teams' chances!
The Division I NCAA women's soccer selection show will take place on Monday at 4:30 p.m. Watch on NCAA.com.
I thought the discussion was whether the regular season winner should be seeded below the team that finished seventh. I think that is a pretty unique situation, but either way, happy to drop it.