Using the Massey ratings (converted to the old Albyn Jones rating scale) here's the expected win probabilities for the 2nd round. I'm using the figure of 60 rating points for homefield advantage. favored team / rating / *homefield / rating diff / expected win prob. / underdog / rating *(H) = favored team at home; (N) = neutral field The rating difference is adjusted for homefield advantage (60 points) Code: Maryland 1764 (N) 32 0.552 Auburn...... 1732 Long Beach St 1753 (N) 60 0.602 Miami...... 1693 Boston College 1778 (N) 67 0.611 California 1711 Penn St...... 1894 (N) 84 0.635 Marquette 1810 Texas A&M 1834 (N) 93 0.651 Virginia Tech 1741 North Carolina 1859 (N) 96 0.659 Baylor.... 1763 Milwaukee 1789 (N) 120 0.697 Ohio St...... 1669 Oklahoma St 1900 (H) 126 0.704 Illinois 1834 Florida .... 1761 (H) 160 0.752 UCF...... 1661 Virginia.... 1877 (H) 220 0.821 Washington St 1717 UCLA...... 1914 (H) 231 0.831 San Diego 1743 Florida St 1883 (N) 233 0.831 Portland 1650 Memphis...... 1880 (H) 261 0.858 Louisville 1679 Wake Forest 1901 (H) 271 0.867 Boston U 1690 Duke...... 1947 (H) 335 0.911 Georgia.... 1672 Stanford 2095 (H) 405 0.943 So. Carolina 1750
Kolabear, I've been meaning to ask: Do Elo systems (does Massey) value recent results more highly than older results? If so, can you tell us anything about how the valuing system works?
Short answer: they can. In chess (where Elo ratings originated): they do. Massey's "Elo-ish" rating (as opposed to the "Power" rating): I don't know Albyn Jones: it did For chess, where the ratings stay with an individual player - it makes sense that it does (as a young player gets better, it doesn't make much sense that the results from when he or she started out weigh the same) When adapted to team sports played within a season - it can but it doesn't have to. You could say it depends on what you're trying to accomplish. If you're trying to more accurately predict the next game, it may make sense to. If you're trying to give an assessment of a team's performance through the entire season, it may make sense not to. If the Committee were to use an Elo system for official purposes, I would argue that it should not weigh recent results more heavily but instead weigh all games throughout the season equally.
I am in the position of seeing Memphis play several times, and not any of the other teams. In my admittedly biased opinion, Memphis is a strong team. Great passing team, and a tremendous offensive threat in Oduro. Anxious to see how they show against Louisville Friday.
Just looking at records you never know about matchups -- Styles Make Fights and you can also always have a bad day. But based on the records, Memphis has beaten teams at Louisville's level - like LSU and UCF. They've beaten teams at the level of Louisville's Big East opponents -- and with nary a hitch (just the lone tie against SMU) I'm reasonably sure Memphis deserves a much higher rating than Louisville The question is whether they're at that next level where the Florida States and Virginias are. Or even close to it. That takes a certain faith in the numbers - it's hard even for me. But I think they're close.
I was considering the Louisville v Memphis/Portland v Florida State group of four next week and thinking, "Wow, that's got to be hard to beat" as a group. Then, I took a look at the other groups of four. What a great weekend coming up. The current format produces great second weekend groups. I can see why the NCAA went to the format.
I did a little twist on Pipsqueak's probability table and thought I'd toss it out there. Like Pipsqueak's table, it's based on the probable opponents that a team will face based on its placement in the bracket -- but using a fixed rating instead of a team's own rating to measure the strength-of-schedule that it is likely to face. It's a way of trying to measure the strength-of-schedule and compare it to other teams. I used 1875 as a rating because that would be approximately the median of the 16 seeded teams if Massey's ratings were used by the Committee to form the bracket. (Numbers converted into the AJ scale again) For the 16 seeded teams here is (was) the probability of their advancing through the 3rd round based only on their opponents' ratings and an assumed rating of 1875* (ranked from highest probability to lowest; you may interpret this as the easiest schedule to the hardest) Code: 1 Duke...... 0.498 1 Stanford 0.417 2 Virginia 0.404 2 Florida.... 0.389 2 Oklahoma St 0.386 3 North Carolina 0.367 1 Wake Forest 0.35 4 Memphis.... 0.339 1 Florida St.... 0.313 2 UCLA...... 0.301 3 Auburn.... 0.269 4 Penn St.... 0.258 3 Texas A&M 0.254 3 Pepperdine 0.233 4 Tennessee 0.193 4 Boston College 0.103 In general, the higher seeds have the easier schedule as measured this way and the lower seeds have the harder schedule as you'd expect. But that's in general - there are several exceptions. * Am I making that clear? This is the probability of a team rated 1875 advancing playing Duke's projected path through the 3rd round, Stanford's projected path, etc.
Did I miss something about who hosts the next rounds? If a higher seed loses then they go to the next seed? Or is all about who offers to host then see?
The #1-8 seeds were, as far as the consensus here seems to think, entitled to hold rounds 2 and 3. Memphis, however, is taking Florida State's place for reasons under discussion by some of us here. The sites for rounds 2 and 3 are set. Even if the host team loses in the 2nd round, the 3rd round will still be played there. As for the 4th round, I think most of us assume the highest remaining seed would be entitled to host the 4th round, assuming they put in a bid to host it. But I don't know if any of us know for sure.
If FSU makes it through rounds 2 & 3 then round 4 will be back in Tallahassee rather than remain in Memphis.. There were to many FSU Athletic events going on in Tallahassee during the coming weekend to accomodate all the expected participants and spectators.
I looked on Top Drawer and all the matches are at 5:00pm. So they just haven't updated the info to reflect three teams going to one site? I'll check NCAA to see unless someone knows better place to go to confirm where teams are going in case Fla State situation occurs elsewhere
So far, every game has been or will be (this coming weekend) at the site of the seeded team or, if there is no seeded team, at the site of the team with the better RPI, with two exceptions. The two exceptions occurred when the team that would have been the host declined to bid to host the game (Dayton and Florida State). This is consistent with what the NCAA has done in prior years. That being the case, I think it's safe to assume that future games will be at the site of the higher seed or, if there is no seed, at the site of the team with the better RPI, so long as the "preferred host" team bids for the game and has a field that meets the NCAA's specifications. Or, are you asking about this coming weekend's games? If so, there are going to be eight sites with four teams at each site. If you check the NCAA printable bracket, for each group of four the site will be the team that has an *.
For the times of all of Friday's games, check Chris Henderson's piece at All White Kit, here: http://www.allwhitekit.com/?p=8278
In the case of FSU . If they winn both games this weekend the following round will revert back to FSU for the quarterfinals.
I created an index showing the relative difficulty of the seeds' projected path through 3 rounds (as determined by Massey ratings converted to the Albyn Jones scale). 100 = hardest schedule, 0 = easiest. Code: 4 Boston College 89 4 Tennessee 75 3 Pepperdine 69 3 Texas A&M 65 4 Penn St.... 65 3 Auburn.... 63 2 UCLA...... 58 1 Florida St 56 4 Memphis.... 52 1 Wake Forest 50 3 North Carolina 47 2 Oklahoma St 45 2 Florida.... 44 2 Virginia.... 42 1 Stanford 40 1 Duke...... 27 In an ideal world, the #4 seeds would have the hardest schedule followed by the #3 seed,then by the #2s and finally the #1s with the easiest schedule. You can see the bracket wasn't set up that way. #1 Florida St (and Wake Forest for that matter) seemingly got a tough draw while #4 Memphis seemingly got a favorable one - for a #4. North Carolina got a favorable draw for a #3, thanks in large part to being in the same group as #2 seed Florida, who was by far the lowest rated #2 seed. In fact, Florida had the lowest rating of any of the seeds, except for #4 seed Auburn.
Memphis may have gotten an easy draw as a #4 seed, but they got a pretty crappy one for a #5 RPI team.
The information is listed: If you just click on the game you're interested in, the date, time and location will pop-up. This information has been on listed since Sunday! If you are interested in a certain game just click on that game while the game is being played and it will give you an update. Similar to gametracker but not as detailed. http://www.ncaa.com/interactive-bracket/soccer-women/d1 Good Luck! OMG1, I heard the same thing about the FSU game!
While this is the way the NCAA basketball brackets are constructed, with every team being "seeded" and placed in a specified bracket line based on the seed, it's not typical of sports brackets generally. Although there no doubt are exceptions, typically, seeds are placed in specified lines of a bracket and the other lines are filled in with non-seeds through a blind draw regardless of how unbalanced the blind draw makes the bracket. Division I women's soccer, on the other hand, doesn't follow either of those procedures at least not from a pure perspective. We know that considerations of travel costs play a significant role in how the bracket is formed. Beyond that, we don't know if the Committee tries to balance the bracket or if it uses more of a blind draw procedure.
Memphis is a tough one to figure out what's fair. In the Massey ratings I think they were 7, 8 or 9. But also when I try to simulate the rating I don't come close -- my calculations put them significantly higher. Perfect records, or nearly perfect ones (Memphis only had one tie keeping them from a perfect regular season record), are a challenge for rating systems...
This is a good point. We should expect the "luck of the draw" to play a factor (although with the NCAA it's more "luck of geography"). But the bigger factor -- and one reason why I made that table -- is a) how the seeds are arranged and b) teams that should be seeds (according to a rating system) aren't seeded or, conversely, teams that shouldn't be seeded are. Penn St, for example, should've been a #2 bracket-seed according to Massey's ratings but instead they have the 5th toughest route to the quarterfinals out of all the seeds. Florida possibly shouldn't have been seeded at all (again according to Massey's ratings) but their presence as a #2 seed is a big reason why North Carolina's path to the quarters seems relatively favorable for a #3 seed.
perfect records or near perfect records would be less of a challenge if the committee were more consistent with how they treated the RPI and perfect or near perfect season. In 2008 Notre Dame had a clean slate for the season. thy beat one of the top 6 or so teams that year, UNC. The rest of their schedule was pretty ordinary, featuring Santa Clara in a down year, Penn state-ditto, and the likes of SMU and a then-woeful Michigan. On that basis, the committee gave them a #1 seed even though four or five teams had a higher RPI. I believe UNC, FSU, Stanford, UCLA, Portland, and FSU all had higher RPI's (not 100% sure on FSU). Looking at the brackets, they were not only a #1 seed, they were THE top seed. I'm not saying that was fair or not, but it did set precedent. This year Memphis had the one tie, so even though it was a no loss season, i suppose you could say that it wasn't a perfect record. you could also make the claim that they didn't beat any of the very top teams. But that's what RPI is supposed to gauge, strength of schedule. If you think the RPI has merit, it's pretty hard to justify downgrading their seed. it might have been different if they actually lost a couple games. Either RPI means something or it doesn't. it seems to me you can't just pick and choose whether it does because a team is a BCS team or something.
It is kind of surprising to see Memphis at #6 in the RPI bumped that far down. I assumed their RPI wasn't as high . . . or can the committee move a #4 seed for cost reasons?
Thanks to cpthomas, I think we can imagine the process that the Committee used - they started looking for wins against teams in the pool of candidates for seeding, wins against teams already designated as seeds, record against common opponents, etc. With Memphis not playing as many teams in this pool, you can see how this procedure works against them. Not that I agree with all of it. I think some of it has merit but, you know me, I think a big part of the problem is the committee's use of RPI or, in this case, paradoxically, its not using it. But, if I may speak in Sphinx-like riddles, that's the fault of using RPI to begin with. Memphis ranks #9 in the Massey but what's important is that they have a rating of 1870. The 17th-ranked team in Massey, Boston College, has a rating of 1778. That's almost 100 points below Memphis. That's a significant amount. A team rated 100 points higher than another would be roughly expected to win 2 out of 3 times. So for Memphis to be placed in a lower seed than any team ranked 17th or lower means overriding a very significant difference in the ratings. By contrast, the difference between the #39th ranked team, BYU with a rating of 1691 and the 47th ranked team, Georgetown, with a rating of 1668 is only 23 points, a very small amount considering the standard margin-of-error and all of that-- and which only corresponds to an expected win percentage of a little over .535 for BYU against Georgetown. This is a difference between ratings and rankings. It's one thing to say one team is ranked, say, 8 places above another. Or that one team was placed (seeded) over another that was ranked 8 places over another. But what does it mean? A rating, however, potentially tells you much more -- and in the examples above, it shows that a difference of 8 places means a lot more between #9 and #17 than it does in bubble territory, #39 vs #47.