2010 RPI

Oh, sure, I wasn't trying to say you could get a #1 seed by gaming the system like this, especially since it doesn't include the adjustments. Obviously, given the way schedules are created years in advance, this scheme would be impractical if not impossible. However, it would suggest that a schedule playing mostly 50-100 teams, and a few teams above 20 might be a useful strategy.

What I thought was most interesting was how Element 2 isn't as close to linear as I would have guessed.
A picture helps here:

Based on what I'm seeing tonight, it appears to me that if the Pilots win out, they will get one of the four #1 seeds in the NCAA Tournament. There's a big current RPI spread between the top four teams -- Stanford, North Carolina, Portland, Notre Dame -- and the next group of teams. Plus, there's a tool on the NC-Soccer website that projects forward to the end of the season what a team's ultimate RPI will be if it wins out. Taking that into consideration as well as who the teams have beaten, the Pilots are looking very good right now if they win out.

Below is a link to the NC-Soccer page that has the tool. Once you get to the page, take a look at the Unadjusted RPI column. Click on the column heading and it will arrange the teams in order of Unadjusted RPI, from worst to best. Click on the heading a second time and it will arrange the teams from best to worst. Once you have them from best to worst, click on the Unadjusted RPI cell of the team for which you want to see a graph showing how their Unadjusted RPI will evolve over the course of the season if they keep winning (based on the assumption their opponents' ratings will not change between now and the end of the season). On the Unadjusted RPI line on the graph, click on the dot at the far right end of the graph. This represents the team's end-of-season Unadjusted RPI. If you then look at the top of the graph, you will see what the team's projected URPI will be. By doing this team-by-team, you can get a rough idea of how the different teams' URPIs will end up if each of them wins out. Add bonus points to those URPIs -- the current points are shown on the page -- and you'll get an even better idea. I'm not anticipating that bonus points will result in any significant shifts in the relative rankings of the teams. It's fun!

Here's the link: http://www.nc-soccer.com/wsoccer/2010/index_arpi

UPSF,

What is the random finish tab on nc-soccer.com?

Is this a prediction on how the reg season will end based on roll or dice based on current win-loss percentage?

gnarly wrote:UPSF,

What is the random finish tab on nc-soccer.com?

Is this a prediction on how the reg season will end based on roll or dice based on current win-loss percentage?

this is one of the least useful but most fun features of GoCourage's website. each time you try the link, you will get a different result in order of finish based on chance game results, which pretty much runs you through all the possible outcomes to the season. I have no idea if they are weighted probabilities, but here's my favorite outcome to date.

I just noticed that outcome this morning and really like it too!

gnarly wrote:UPSF,

What is the random finish tab on nc-soccer.com?

Is this a prediction on how the reg season will end based on roll or dice based on current win-loss percentage?

It's actually a neat device that has more to it than one at first might think. It uses the game results for games already played and then inserts an outcome for games not yet played. It uses the teams' current URPI ranking as the determining factor for the outcomes of about 73% of the remaining games (the team with the higher URPI ranking is given a win). For the other 27%: 10% randomly are ties regardless of URPI ranking; 9% randomly are won by the home team regardless of URPI ranking; and 8% randomly are won by the away team regardless of URPI ranking.

The "more to it" comes from the fact that 73% is just about the right number for games won by the team with the higher URPI ranking. So, the device is a fairly rough, but fun attempt to show reasonably possible outcomes for the season.

I thought the predicted results for OSU were generous since they have UCLA this weekend and they also the bay area schools to play. If they can go 500 the rest of the season seems more reasonable.

UPSoccerFanatic wrote:click on the Unadjusted RPI cell of the team for which you want to see a graph showing how their Unadjusted RPI will evolve over the course of the season if they keep winning (based on the assumption their opponents' ratings will not change between now and the end of the season). On the Unadjusted RPI line on the graph, click on the dot at the far right end of the graph. This represents the team's end-of-season Unadjusted RPI. If you then look at the top of the graph, you will see what the team's projected URPI will be. By doing this team-by-team, you can get a rough idea of how the different teams' URPIs will end up if each of them wins out. Add bonus points to those URPIs -- the current points are shown on the page -- and you'll get an even better idea. I'm not anticipating that bonus points will result in any significant shifts in the relative rankings of the teams. It's fun!

Here's the link: http://www.nc-soccer.com/wsoccer/2010/index_arpi

I note that Oklahoma State is projected to overtake Notre Dame and win a #1 seed if both win out. ND should get bonus points for beating Georgetown, and OKS would get bonus points twice if they beat Texas and Texas A&M, so bonus points will not save ND if both teams win out. Then they both have conference tourneys, which the projections do not take into account, but I'm guessing that OKS would still get the seed if they win, since A&M's win-loss record is a little better than Marquette's, etc.

That Oklahoma STate goalie was amazing, and came within minutes of taking us to overtime.

Since there's been some discussion on PN about the probabilities of Portland getting a #1 seed and hosting in the quarters, I thought I'd insert this tidbit of information here:

Last year, as of this time in the season, the Pilots were ranked #4 in the RPI. The difference between the Unadjusted RPI rating of the Pilots and the #5 team (UCLA) was 0.002. As of yesterday, the Pilots were ranked #2 in the RPI, with #3 North Carolina close behind. Even if the Pilots win out, if North Carolina wins the ACC Tournament, they'll probably move up to #2 and drop the Pilots to #3. The distance between the Pilots and the current #4 team (Boston College), however, is 0.045, followed by #5 Notre Dame and #6 Virginia both at a distance of 0.047. You'll note that the 0.045 difference is much larger than last year's 0.002 between the Pilots and UCLA. In RPI terms, 0.045 is a big difference.

This gives some reason for optimism, although as we all know we never can count on anything. It's looking to me like it'll be Stanford and Portland getting #1 seeds, North Carolina getting a #1 unless they encounter a disaster between now and the end of the season, and a battle among Boston College, Virginia, and Notre Dame for the fourth #1.

Looking at the NC-soccer spreadsheet I analyzed the top RPI teams (adjusted tab). I've noted their wins over top 40 rpi teams as listed on the schedule page for the team, and also noted their ties and losses. (The numbers don't exactly tie because that tab uses the unadjusted rpi I believe. This can start to help us get an idea on seeds. For example, I think it was PSU had rpi of 9 a few years ago but did not get a seed. Last year I think MD had an rpi of 20 or so but got a seed.

Stanford 1: Losses:none, Ties BC 4, UNC 3. Wins: GT 16, UP 2, SCU 18, Duke 26, USC 24, UCLA 22, UW 35, ASU 13.

UP 2: Losses: Stan 1, Wins: OK st 12, Wis 23, A&M 9, UW 35, SCU 18.

UNC 3: Losses: BC4, MD 14. Ties: Stan 1. Wins: A&M 9, FL 7, VT 37, Vir 5, FSU 11, and Duke 26. Still has WF 30 to play in reg season.

BC 4: losses: Vir 5, MD 14, VT 37, Duke 26. Ties Stan 1. Wins: Hofstra 15, UNC 3, WF 30

To be continued, have to surrender laptop for kids homework.

Vir 5: Losses: WV 21, MD 14, UNC 3. Ties: FSU 11. Wins Texas 32, VT 37, BC 4, Duke 26, WF 30.

ND 6: Losses UCLA 22, Ties GT 16, wins Minn 34, SCU 18, TT 40.

Ohio St: Losses: Mil 56, ILL 36, ASU 13. Ties MQ 10. Wins: WV 21, Day 27, Wis 23, Minn 34.

A&M 8: Losses: UNC 3, UP 2, Col 89. wins: VT 37, Neb 39, TT 40, Ok st 12. Still has Texas to play.

FL 9: Losses UNC 3, Aub 52. Ties: Duke 26. AL 73. Wins: UCF 17, FSU 11

MD: Losses VT 37, WF 30. Ties Duke 26. Wins: Vir 5, BC 4, UNC 3 and FSU 11.

MQ 11: Losses NM 28, SD 51, BYU 33. Ties MIL 56, Oh St 8. Wins: WV 21 and GT 16.

OK St 12: Losses UP 2, Tex 32, and A&M9. Ties: Col C. Wins: Mem 38, TT 40 and Neb 39.

ASU 13: Losses: UCF 17, LB st 54, UW 35, Stan 1. Ties S. Caro 20, Dep 74, Cal 45. Wins Neb 39, Oh St 8. Still have OSU, USC and UCLA to play.

FSU 14: Losses Aub 52, FL 7, UNC 3, MD 14. Ties Vir 5. Wins: UCF 17, WF 30, Duke 26, and VT 37. Still has BC.

Hofstra 15: Loss BC 4, Best win is #58 Columbia

UCF 16: Losses FL 7, FSU 11, Col C. 67. Ties AZ 105, Mem 38, Rice 80. Wins Asu 13

GT 17: Losses: SCU 18, Stan 1, Rut 70, MQ 10, WV 21. Ties: ND 6. Best win only S. FL 46

UCLA 18: Losses: NW 85, Col 89, Pep 74, Stan 1, OSU 29, USC 24. Ties UCSB 133. Wins Wis 23, ND 23, and SCU 18.

UC-Irv 19: Losses Seattle 81. Ties Pep 74, and Cal 45. Win: OSU 29.

Wis 20: Losses: UCLA 22, UP 2, Cin 97, Oh st 8. Ties : UW 35, SD St 139, Pur 64m Ill 36. Wins Minn 34.

WV 21: Losses: PSU 61, Oh st 8, Miami 42, and MQ 10. Ties S. FL 46. Wins GT 16.

SCU 22: Losses: ND 6, Stan 1, UP 2, UCLA 22. Ties Cal 45 and Geo 41. Wins GT 16

Make your own conclusions: What I see is MD has wins over UNC, VIR and BC and FSU (I could see them moving up to a 2 seed). Good chance to see ASU drop with OSU, USC and UCLA still on their schedule. Hofstra hasn't beat anyone in top 40. Will they get seeded?

The playoff pod picture still cloudy with so many big regular season games left this weekend ( particularly in PAC 10 and Big 10) and with all the conf tourneys this week that will have only one auto bid.

Here is my current thinking on the West Coast. Like I said before LA is likely not to have any ranked high enough to seed. USC has outside chance. If OSU beats Cal this weekend I think they have shot at a seed. If they do, where might they go: LA. They join UC-Irv and either LMU or SD. USC and UCLA get shipped back East somewhere if it works out this way. My thinking is OSU and USC are competing for a seed, OSU gets it based on head to head result.

I was thinking AZ/NM would have a pod. ASU and NM have dropped and they are no longer likely to have a pod there. My thinking is that NM heads to OK St or A&M. N. Az (if they get Big sky bid) play somewhere on West Coast. Maybe LA this year. They went to Bay area last year and UP the year before. ASU heads east IF they make the tourney. Their RPI has been dropping fast.

UPSoccerFanatic wrote:
This gives some reason for optimism, although as we all know we never can count on anything. It's looking to me like it'll be Stanford and Portland getting #1 seeds, North Carolina getting a #1 unless they encounter a disaster between now and the end of the season, and a battle among Boston College, Virginia, and Notre Dame for the fourth #1.

Notre Dame? you mean those golden dome folks? No way.

Based on what the committee said last year about UP not having played enough top 25 teams to get a #1 seed, I just don't see how Notre Dame is in the running this year, especially since they are sitting home for their tournament semis and final.

in the top 20 they beat SCU at 17 and tied Georgetown at 15. They lost to UCLA at 27.

then their opponents jump to 36 which they beat, and tied and then lost to #38.


Then it's 48, 56, 58, 61 and up from there.

That's absolutely not #1 seed territory. Either OSU has a better claim to a one seed. (not that they have a chance, either)

Geezaldinho wrote:

UPSoccerFanatic wrote:
This gives some reason for optimism, although as we all know we never can count on anything. It's looking to me like it'll be Stanford and Portland getting #1 seeds, North Carolina getting a #1 unless they encounter a disaster between now and the end of the season, and a battle among Boston College, Virginia, and Notre Dame for the fourth #1.

Notre Dame? you mean those golden dome folks? No way.

Based on what the committee said last year about UP not having played enough top 25 teams to get a #1 seed, I just don't see how Notre Dame is in the running this year, especially since they are sitting home for their tournament semis and final.

in the top 20 they beat SCU at 17 and tied Georgetown at 15. They lost to UCLA at 27.

then their opponents jump to 36 which they beat, and tied and then lost to #38.


Then it's 48, 56, 58, 61 and up from there.

That's absolutely not #1 seed territory. Either OSU has a better claim to a one seed. (not that they have a chance, either)

Well, when I wrote the post (October 26), it looked like Notre Dame could get the fourth #1 seed. Now, with their loss to Connecticut in the Big East tournament, there's no way.

It still looks like Stanford, Portland, and North Carolina at the top. Then one of Boston College, Oklahoma State, Virginia, or Maryland. Something I don't like is not knowing how the Women's Soccer Committee will look at the Pilots' loss to Stanford @ Stanford, as compared to Stanford's ties against Boston College @ Boston College and against North Carolina @ North Carolina. We don't want that to work against us, which it shouldn't.

The main thing though, is for the Pilots to keep their focus narrow, on the game this weekend against St. Mary's. That is the only thing that matters right now.

Geezaldinho wrote:

eProf wrote:

UPSoccerFanatic wrote:On the other hand, over the period for which NCAA Tournament brackets are available, UNC never has played an away game -- its only non-home games have been at the "neutral" College Cup site.

I have a vague memory of a year when UNC was seeded 5th or 6th overall and had an away playoff game. In Nebraska around 1997, maybe?

UNC, unlike (ahem) some sites has records going back that far. They show 27 wins in 1997-1998, so that couldn't have been it. they were 17-0-1 during the regular season. they played 28 that year, which may be a record.

They also show 25 wins for 1992-1994, 1995 and 1996, so I suspect they weren't a 5-6 seed then, either.

They won it all in 1999.

UNC was the #5 overall seed in 2000 (behind Notre Dame, Washington, Clemson, Nebraska). Nebraska, however, lost to UConn in the round of 16 that year, so UNC played the quarterfinal at home.