NCAA Tournament Bracket Formation

Here is my first run at what the NCAA Tournament roster would be if the NCAA had to decide on it today (from the BigSoccer blog):

Using the NCAA's most recently released RPI report, and then using head to head results and results against common opponents, I've looked at who I think would get at large berths. It looks to me like the range of potential conference champions coming from the top 50 RPI-ranked teams is 8 to 16. Eight seems pretty certain and 16 probably is high. Based on my review, I don't see head to head results and results against common opponents providing a significant reason to deviate from the RPI rankings (with possibly one exception). I have not yet addressed possible seeding issues. Here are the teams in order based on the NCAA's RPI:

1. UCLA

2. North Carolina

3. Stanford

4. Florida State

5. Portland

6. Notre Dame

7. Florida

8. Boston College

9. Texas A&M

10. Duke

11. Viriginia

12. Penn State

13. Oklahoma State

14. USC

15. Wake Forest

16. Texas

17. Colorado

18. Washington

19. BYU

20. Long Beach State

21. Milwaukee

22. Minnesota

23. San Diego

24. UCF

25. Princeton

26. Miami FL

27. Illinois

28. Missouri

29. Virginia Tech

30. Rutgers

31. West Virginia

32. South Carolina

33. LSU

34. Charlotte

35. Denver

36. Michigan State

37. William & Mary (This is the one team that I see could be dropped down based on head-to-head results. They lost to Columbia (49) and James Madison (45) and tied Washington State (44). I could see dropping them to 45 and moving the current 38 to 45 up one position.)

38. Auburn

39. California

40. Kansas

41. Harvard

42. Marquette

************* Cutoff if only 8 of 16 potential teams are automatic qualifiers.

43. East Carolina (if 9 are AQs)

44. Washington State (if 10 are AQs) [Note: This is one where I think the issue of the RPI's problems with cross-regional rankings comes into play. WSU has tied USC (14), Washington (18), William and Mary (37), and James Madison (44);and beat Cal (40). If there's doubt between WSU and Wiliam and Mary, for example, my thought is that WSU gets the nod because of the likelihood the RPI somewhat under-rates West Region teams.]

45. James Madison (if 11 are AQs).

46. Arizona State (if 12 are AQs). [Note: See the WSU note. Arizona State tied Texas (16) and Cal (40). It beat USC (14) and Long Beach State (20). But, it lost to TCU (50) and Oregon (51).]

47. Georgia (if 13 are AQs and if Georgia's record is either .500 or better or it wins the SEC Tournament. If Georgia doesn't win the SEC Tournament and its record is below .500, then it cannot get an at large berth.)

48. UNC Greensboro (if 14 are AQs)

49. Columbia (if 15 are AQs)

50. TCU (if 16 are AQs)

51. Oregon (if 16 are AQs, Georgia doesn't win the SEC Tournament, and Georgia's record is below .500)

Notre Dame?

Thanks, Geez. I had North Carolina twice, once at #2 (where the RPI has them), and once at #6 (in Notre Dame's RPI spot). I edited the list to fix it.

I didn't catch that UNC was twice. I just knew there was an imbalance in the world order...

Looks like WSU is on the bubble. It obviously would be huge if they could move up or a few others could get in.

I haven't run any numbers, but I'm thinking that WSU needs a tie and a win to be safe about getting into the tournament. Oregon might be able to get in with a tie -- or, maybe, even with two losses.

As an aside, I've been corresponding with one of the WSU assistant coaches about RPI issues. He says he coached Kassie when she played club for Soreno and also Ellen Parker. He says they knew Kassie was going to be really good and thought she would thrive with Garrett as a coach. He says Ellen will make a great addition to our midfield.

Later addition:

I changed my mind and ran the numbers, using the following hypothetical:

Washington wins against Arizona and Arizona State

Washington State ties against one Arizona and wins against the other (it doesn't matter which is which)

Oregon loses both games against the So Cal teams.

Here are the results:

Washington: Its unadjusted RPI rank moves up from 16 to 14. Last week, its adjusted RPI rank was two spots below its unadjusted rank. If this holds, Washington's adjusted RPI rank would be at 16. Absent contrary head to head or common opponent results, they would get a #16 seed. A tie or loss probably would put take them out of the seeding area.

Washington State: its unadjusted RPI rank moves up from 48 to 47. Last week, its adjusted RPI rank was five spots above its unadjusted rank. If this holds, WSU's adjusted RPI rank would be at 42. There don't appear to be head to head or common opponent results to move them down and might be results to move them up. So, they almost for sure would get in. I have not run numbers to see where they fall with a win and a loss.

Oregon: Here's the interesting one. Notwithstanding two losses in this hypothetical, their unadjusted RPI rank moves from 51 to 48. This is due to the improvement to their strength of schedule. Last week, Oregon's adjusted RPI rank was three spots above its unadjusted rank. If this holds, Oregon's adjusted RPI rank would be 45. There is one negative head to head result against a slightly lower-ranked opponent, Purdue. If Purdue were jumped up above Oregon, then Oregon would be 46. Either 45 or 46 should have a very good chance of getting an at large selection.

Remember, however, that these hypotheticals do not consider non-conference opponents' results this week or the results of all the other teams in contention. So, take them with a grain of salt.

Sorry to make you explain this again, UPSF, but I'm puzzled that you saw no head-to-head or common-opponent reason to deviate much from RPI.
How about moving Notre Dame up and NC down because of their head-to-head? How about switching UP and UF, who are so close in RPI, because of their results against common opponent Florida?

I just wish we could get that #4 seed (or higher) so that we wouldn't meet UCLA until the College Cup.

And I get creeped out by the system's worship of strength-of-schedule over all else. It's outrageous that NC is so much higher than ND when ND not only beat NC but has a much better record. Stanford and UP also have better records than NC. It would be very difficult (and expensive, and disruptive of education) for UP to beef up their strength-of-schedule any stronger than it was this season, without moving to a different conference.

OK DWM, If you want to move ND up and UNC down, don't you have to go through UP to do it?

How do you justify putting ND above UP when there is no linkage between them that differentiates the teams?

Actually, we beat Florida, who beat FSU, who tied NC, who lost to ND. Does that count as a linkage?

I was wrong about there being linkage between ND and UP. I think you are allowed to compare head-to-head and common opponents.

We didn't play them, so the first part is out.

We played LMU at home and beat them 3-0
ND played LMU at home and beat them 4-0

We played Santa Clara at home and beat them 4-0
ND played Santa Clara at home and beat them 2-0

We played Penn State away and beat them 4-1
ND played Penn State away and beat them 3-1

We played Rutgers away and beat them 1-0
ND played Rutgers at home and beat them 3-1

The results were the same
We have the better goal differential (11-9)
we had an extra away game

Slam dunk - we win

Sorry, I mis-communicated. In referring to no head to head or common opponent results, I only was referring to Washington State's position. I haven't done a detailed look yet at seeds. However, regarding the Portland to Florida to Florida State-North Carolina-Stanford (linked by common ties) to Notre Dame (they beat North Carolina) linkage, a question regarding North Carolina and Notre Dame is this: if we allow Notre Dame's head to head win over North Carolina to convince us they are better than North Carolina notwithstanding the RPI, what do we do? Do we jump Notre Dame up the rankings until it is ahead of North Carolina-Stanford? Or, do we drop North Carolina-Stanford down through the rankings until they are below Notre Dame? For the Pilots, it makes a big difference what we do

I don't think it is permissible to link opponents of opponents of opponents of opponents.

The seeding rules only talk about head to head and common opponents. The RPI already made the deeper links, according to the NCAA's logic . If you allowed the endless chain you are just repeating the RPI results.

Here's what they allow.

Primary Criteria in priority order
• Adjusted Rating Percentage Index (RPI);
(See Appendix I of the 2008 NCAA Division I Women’s Soccer Championship Hand-
book).
• Results versus common opponents; and
• Head-to-head competition.

Secondary Criteria
If the evaluation of the primary criteria does not result in a decision, the secondary
criteria will be reviewed. All the criteria listed will be evaluated.
• Results versus teams already selected to participate in the field (including
automatic qualifiers with and RPI of 1-75)
• Late season performance - defined as the last eight games including
conference tournaments (strength and results).

In response to a question someone asked somewhere about who I think is our best chance to host rounds 1/2, of Washington, Washington State, Oregon, and Portland State, I'll go out on a limb and give my best guess. FANatic, if I had an emoticon of a happy face sawing off a tree limb while sitting on the wrong side of it, I'd insert it here.

Not Washington. I'm guessing they'll get a seed.

Possibly Washington State, but I still think they need a tie and win in their two Arizona games. However, their stock has improved since yesterday based on today's results. Sort of weird, but that's the way the RPI works.

Portland State is a total question mark. I'd love them to make it, especially for Coach Laura Schott from Jesuit via Cal, but I think that's not something I would bet money on.

Oregon is my choice, even if they lose both games. UCLA and USC will do wonders for their strength of schedule. The only danger facing them is that Purdue is ranked lower than they are and Purdue beat them. So, Purdue might advance ahead of them. Even if that happens, I think they've got a good shot. A tie in one of their two games and I think they're even better off.

Do we have any determination of what the phrase "in priority order" in the excert in my last post means?

If the first priority is RPI, how do lesser priorities trump that?

Geezaldinho wrote:I don't think it is permissible to link opponents of opponents of opponents of opponents.

The seeding rules only talk about head to head and common opponents. The RPI already made the deeper links, according to the NCAA's logic . If you allowed the endless chain you are just repeating the RPI results.

[/quote]

I agree with this. However, here's where I have a question about what this means. A and B tie. C beats B. D beats C. The RPI order, however, is A, B, C, D. C argues that it should be ahead of B by virtue of head to results. C also argues that it should be ahead of A by virtue of results against a common opponent. Let's say we agree with C. D then argues it should be ahead of C by virtue of head to results. Let's say we also agree with D. By this series of decisions, D gets ahead of A. This series of steps is literally consistent with what you said and with the rules.

So, applying that logic to our situation, since Florida beat Florida State, it could be placed ahead of FSU by virtue of head to head results. Further, since Florida State tied North Carolina, Florida could be placed ahead of North Carolina by virtue of results against a common opponent. So, fine, the new order is Florida-North Carolina-Florida State. Portland beat Florida, so it is entitled to be ahead of Florida by virtue of head to head results as well as the RPI.

Geez, do you think the rules prohibit that line of reasoning?

[quote="UPSoccerFanatic"]

Geezaldinho wrote:
Geez, do you think the rules prohibit that line of reasoning?

The way I read it, the "in priority order" phrase is key. It seems to me that means if you have a clear determination by RPI you are done.

I don't see how a lesser priority can override the greater priority. Perhaps if both lesser priorities overrode, I could see that you could reverse the rpi determination, but the rule isn't even clear on that.

Geezaldinho wrote:Do we have any determination of what the phrase "in priority order" in the excert in my last post means?

If the first priority is RPI, how do lesser priorities trump that?

I'm looking right now at the 2008 Division 1 Women's Soccer Championship Handbook, page 10 (updated 9/15/2008). It says:

"Selection Criteria

"The following criteria shall be employed by a governing sports committee in selecting participants for NCAA championships competition [Bylaw 31.3.3; Criteria for Selection of Participants]:

"* Won-lost record;

"* Strength of schedule; and

"* Eligibility and availability of student-athletes for NCAA championships [which I think refers to academic eligibility, etc., etc.];

"In additionto Bylaw 31.3.3, the women's soccer committee has received approval from the Division 1 Championships/Competition Cabinet [recently reconstituted under a different name] to consider the following criteria in the selection of at large teams for the soccer championship (not necessarily in priority order:

"Primary Criteria

"* Adjusted Rating Percentage Index (RPI); (See Appendix I of the 2008 NCAA Division 1 Omen's Soccer Championship Handbook).

"* Results versus common opponents; and

"* Head-to-head competition."

"Secondary Criteria

"If the evaluation of the primary criteria does not result in a decision the secondary criteria will be reviewed. All the criteria listed will be evaluated.

"* Results versus teams already selected to participate in the field (including automatic qualifiers with an RPI of 1-75)

"* Late season performance -- defined as the last eight games including conference tournaments (strength and results)."

I'm not sure where your quote "in priority order" comes from, but it seems directly to what the Manual says.

Hmm.. that's different than the one I'm reading. I must have an older version of the 2008 handbook.


OK, accepting your version, I still see problems:

Looking at your A,B,C,D example again, the linkage is broken between D and A because A didn't play C. Therefore it is not permitted under the rules to use C in the common opponent link.

C can't make a claim against A based only on the common opponent link
since A and C didn't meet, it seems clear that there is no third comparison without a head to head to break the tie, so it appears we go to secondary criteria.


So it's comparison against other teams in the tournament (1-75) and last 8 games.

I assume that means won-loss-tie record against those top 75 teams getting into the tournament?

But since what you are doing is CREATING the tournament field, how do you determine the proper teams to use for comparison?

That logic could possibly be defended IF all the teams had RPIs and records very close to each other. However, both Floridas are significantly below the top five teams, since they've each lost twice and tied at least twice.

As for what I'd do, I'd put
ND over UCLA (on basis of records, and common opp Santa Clara);
UCLA over Stanford (records, and head to head)
Stanford over UP (very close, but RPI, and somewhat better scores against common opps)
UP over NC (very close again; records, and linkage via the 2 Floridas, and I don't spot any strong linkage to warrant NC ranking above any of those top 3)
NC over the Floridas (RPI, records)

So that's my top 5, but what the hell. I just checked the Massey and Jones ratings, and if you throw in the coaches ratings and the RPI, you get those top 5 teams in almost any order. Massey still has Stanford at #2 somehow, after ND; I'm curious as to what mechanism could put Stanford above UCLA at this point. Jones has UP at #3, below ND and NC but above UCLA.

But back to reality. Are the above-quoted rules, involving head-to-head and common opps, supposed to be used only in deciding at-large entries? Or also in deciding seeds? I thought I once read (doubtless in a post by UDSF) that the "adjustments" to RPI are bonus and penalty points for remarkable upsets achieved by away teams? Well, I guess I haven't seen any of those upsets affecting the top five teams. UCLA was tied, not beaten, and neither tie was at home in LA.

dwm wrote:
...


As for what I'd do, I'd put
...


Stanford over UP (very close, but RPI, and somewhat better scores against common opps)
UP over NC (very close again; records, and linkage via the 2 Floridas, and I don't spot any strong linkage to warrant NC ranking above any of those top 3)
NC over the Floridas (RPI, records)

.

Here's how I see it
The rule talks about results, not scores, so you can't have Stanford over UP based solely on the score.

Since they didn't meet, you have to go to secondary criteria.
There, UP has a better record against the tournament teams in the top 75, since Stanford tied UNC

Also, they have a loss in the last 8 games, where UP doesn't

Therefore, UP is ranked over Stanford.

I'll do UNC next

The only common opponent I see between UNC and Portland is Santa Clara. Both schools won. UNC has the higher RPI, so we welcome UNC in the #1 seed club.

dwm wrote:
As for what I'd do, I'd put
ND over UCLA (on basis of records, and common opp Santa Clara);
.

The criteria isn't record, it's RPI , common opponents and head to head.

UCLA wins on the RPI, ND wins on the common opponents

That one goes to secondary criteria

They have both won their last 8 games, so I guess it goes t tournament teams in the top 75.


You can't do that until the field is selected. How many tournament teams in the top 75 did ND play? (Same question for UCLA)

Hmm.. Forget all that stuff, I guess.

As I read the rules as just posted by UPSF, there is nothing there that pertains to seeding.

DWM is right. All those criteria are for

the selection of at large teams for the soccer championship

So It's not only not for for seeding but for selection, but a strict reading says it isn't to be used with the automatic qualifiers at all.

They don't have any published seeding criteria that I see. They do talk about Pairings, and say the top 16 teams will be seeded. but they don't say how the top 16 teams will be determined.

Pairings
[Reference: Bylaw 31.1.3 in the NCAA Division I Manual.]
The top 16 teams identified by the committee will be seeded in the bracket. The
seeded teams will compete at separate first and second round sites. The committee
has been given approval by the Championships/Sports Management Cabinet to place
the top 16 seeds in pods of four. There will be four number ones, four number twos,
four number threes, and four number fours. The remaining 48 teams will be paired
geographically. First- and second-round conference matchups will be avoided. The
committee will use the NCAA’s mileage calculator located on the travel Web site at
http://www1.ncaa.org/finance/travelindex.html. when establishing the mileage for
travel. Teams within 400 miles of each other will be required to drive.

Near as I can tell, Seeding is whatever the committee says it is.

How about them apples?

uuuhhhhh....... I just woke up. Did I miss the Bracket, RPI, Seeding lecture by the math professors?

(Or is it the 'statistical physics' professors?)

How of a drop will FSU make after yesterdays utter failure. Do you think we can still get a 2 seed? Also should we be cheering for Virginia to win the ACC Tournament, since they are the only ACC team to beat us during the regular season?