I'm in the process of building a 2010 master schedule for RPI purposes and as a sidelight thought I'd evaluate at least the upper group of teams' 2010 strengths of schedule. As the measure of opponents' strengths, I'm using opponents' 2009 post-NCAA Tournament adjusted RPI ratings. Since I use my best guess as to the adjustment bonuses and penalties, the ratings aren't exactly correct, but they're very close. I've done the evaluation so far for the top 10 teams in the ARPI post-Tournament ratings, except that neither Florida State nor Virginia Tech has posted its schedule on-line, so I can't evaluate their schedules yet. Here are the results for the top 10 teams, with information in the following order: rank of team in strength of 2010 schedule, team, team's 2009 post-Tournament ARPI rank, and the average 2009 post-Tournament rating for the team's 2010 opponents: 1. North Carolina (1) .6063 1. Stanford (2) .6063 3. UCLA (3) .5959 4. Wake Forest (8) .5933 5. Boston College (7) .5918 6. Portland (4) .5746 7. Notre Dame (5) .5604 8. South Carolina (9) .5589 NA: Florida State (6) and Virginia Tech (10)
Here is a new list with teams 11-20 added: 1. North Carolina (1) .6063 1. Stanford (2) .6063 3. UCLA (3) .5959 4. Wake Forest (8) .5933 5. Boston College (7) .5918 6. Washington State (18) .5863 7. Penn State (16) .5761 8. Portland (4) .5746 9. Georgia (20) .5743 10. Maryland (14) .5739 11. Notre Dame (5) .5604 12. South Carolina (9) .5589 13. LSU (13) .5568 14. UCF (11) .5451 Not Available: Florida State (6), Virginia Tech (10), Santa Clara (12), Florida (15), USC (17), Oregon State (19)
The ACC and Pac-10 are the toughest conferences so it's not a surprise their top teams would also have the toughest nonconference schedules, too.
And, here's an update through the top 30 2009 ARPI teams. For an explanation of my method, go to the first post on this thread. The opponents' average 2009 ARPIs include both conference and non-conference opponents. 1. North Carolina (1) .6063 (ACC) 1. Stanford (2) .6063 (Pac 10) 3. UCLA (3) .5959 (Pac 10) 4. Wake Forest (8) .5933 (ACC) 5. Boston College (7) .5918 (ACC) 6. Washington State (18) .5863 (Pac 10) 7. California (30) .5836 (this year, 14 home games, 5 away) (Pac 10) 8. Washington (26) .5763 (Pac 10) 9. Penn State (16) .5761 (Big 10) 10. Portland (4) .5746 (West Coast) 11. Georgia (20) .5743 (SEC) 12. Maryland (14) .5739 (ACC) 13. Texas A&M (22) .5707 (Big 12) 14. Ohio State (21) .5607 (Big 10) 15. Notre Dame (5) .5604 (Big East) 16. South Carolina (9) .5589 (SEC) 17. LSU (13) .5568 (SEC) 18. West Virginia (25) .5542 (Big East) 19. UCF (11) .5451 (CUSA) 20. BYU (28) .5350 (Mountain West) Not available: Florida State (6), Virginia Tech (10), Santa Clara (12), Florida (15), USC (17), Oregon State (19), Marquette (23), Virginia (24), Connecticut (27), San Diego State (29).
Funny, I wondered when the first post about the 2010 season schedules was going to happen. I've been working on my version of the 2010 schedule in parallel. The results are on my website. By volume I think I have about 2/3 of the games that should be scheduled. Each of the team pages breaks down the opponents, their opponents records (average winning percentage), and the 'opponents strength'... which is calculated as 1/4 of the element 2 (*2) and element 3 (*1)... from your previous end of season spreadsheets (Woohoo for open information, cpthomas! ). If you take the OS and 1/4 of a teams predicted winning percentage it should give a reasonable guess at the relative RPI value. The website also gives previous seasons if I have them available so you can possibly see trends for opponents. Example page for UNC: http://www.nc-soccer.com/wsoccer/2010/schedule/1
GoCourage, I thought you might like to know that your site appears to be fully functional on an iPad with Safari, including all the Java windows and links. Hot Diggety.
Check out GoCourage's 2010 data pages, everyone, they're great. The link is two posts previous to this one. I've now got schedules for 47 out of last year's top 60 teams in the ARPI rankings. Here's how they stack up in terms of their 2010 opponents' average strengths, with strength measured by the team's final post-NCAA Tournament 2009 ARPIs. I've got both the average strength of all of each team's opponents and the average strength of each team's non-conference opponents. For those interested in how teams schedule non-conference opponents, this provides some interesting fodder. I'll run two separate lists. The first is in order of strength of all opponents. The second is in order of strength of non-conference opponents. The order of the data is Team -- 2009 ARPI Rank -- Average 2009 ARPI of 2010 Opponents -- Average 2009 ARPI of 2010 Non-Conference Opponents. The delinquent schools with no schedules posted yet are Florida State (6), Virginia Tech (10), Santa Clara (12), Florida (15), Oregon State (19), Marquette (23), Virginia (24), Connecticut (27), Rutgers (39), Missouri (44), San Diego (47), Michigan State (50), and Samford (54). Sorted by Average Strength of All 2010 Opponents: North Carolina U 1 0.6063 0.6030 Stanford 2 0.6063 0.6023 Arizona State 32 0.5981 0.5690 Duke 43 0.5969 0.5616 UCLA 3 0.5959 0.5779 Wake Forest 8 0.5933 0.5626 Boston College 7 0.5918 0.5565 Oregon U 58 0.5864 0.5473 Washington State 18 0.5863 0.5531 Miami FL 49 0.5837 0.5333 California U 30 0.5836 0.5420 Texas A&M 22 0.5774 0.6162 Washington U 26 0.5763 0.5332 Penn State 16 0.5761 0.5812 USC 17 0.5756 0.5336 Portland U 4 0.5746 0.5864 Georgia U 20 0.5743 0.5733 Maryland U 14 0.5739 0.5125 Wisconsin U 31 0.5721 0.5693 Auburn 40 0.5706 0.5596 Vanderbilt 51 0.5661 0.5468 Minnesota U 37 0.5640 0.5519 Villanova 45 0.5628 0.5594 Purdue 38 0.5616 0.5471 Ohio State 21 0.5607 0.5471 Notre Dame 5 0.5604 0.5859 South Carolina U 9 0.5589 0.5418 Georgetown 56 0.5588 0.5471 LSU 13 0.5568 0.5330 West Virginia U 25 0.5542 0.5550 St Johns 41 0.5526 0.5243 Mississippi U 35 0.5520 0.5199 San Diego State 29 0.5496 0.5702 Indiana U 53 0.5467 0.5165 Oklahoma State 36 0.5461 0.5468 UCF 11 0.5451 0.5742 BYU 28 0.5350 0.5567 Memphis 34 0.5338 0.5450 UAB 59 0.5328 0.5390 Colorado College 42 0.5286 0.5312 Kansas U 55 0.5275 0.5050 UNC Wilmington 57 0.5103 0.5073 UNC Greensboro 60 0.5041 0.5646 Northeastern 52 0.4931 0.4508 Charlotte 46 0.4918 0.5270 Dayton 33 0.4733 0.4987 Central Michigan 48 0.4687 0.4670 Sorted by Average Strength of 2010 Non-Conference Opponents: Texas A&M 22 0.5774 0.6162 North Carolina U 1 0.6063 0.6030 Stanford 2 0.6063 0.6023 Portland U 4 0.5746 0.5864 Notre Dame 5 0.5604 0.5859 Penn State 16 0.5761 0.5812 UCLA 3 0.5959 0.5779 UCF 11 0.5451 0.5742 Georgia U 20 0.5743 0.5733 San Diego State 29 0.5496 0.5702 Wisconsin U 31 0.5721 0.5693 Arizona State 32 0.5981 0.5690 UNC Greensboro 60 0.5041 0.5646 Wake Forest 8 0.5933 0.5626 Duke 43 0.5969 0.5616 Auburn 40 0.5706 0.5596 Villanova 45 0.5628 0.5594 BYU 28 0.5350 0.5567 Boston College 7 0.5918 0.5565 West Virginia U 25 0.5542 0.5550 Washington State 18 0.5863 0.5531 Minnesota U 37 0.5640 0.5519 Oregon U 58 0.5864 0.5473 Ohio State 21 0.5607 0.5471 Purdue 38 0.5616 0.5471 Georgetown 56 0.5588 0.5471 Vanderbilt 51 0.5661 0.5468 Oklahoma State 36 0.5461 0.5468 Memphis 34 0.5338 0.5450 California U 30 0.5836 0.5420 South Carolina U 9 0.5589 0.5418 UAB 59 0.5328 0.5390 USC 17 0.5756 0.5336 Miami FL 49 0.5837 0.5333 Washington U 26 0.5763 0.5332 LSU 13 0.5568 0.5330 Colorado College 42 0.5286 0.5312 Charlotte 46 0.4918 0.5270 St Johns 41 0.5526 0.5243 Mississippi U 35 0.5520 0.5199 Indiana U 53 0.5467 0.5165 Maryland U 14 0.5739 0.5125 UNC Wilmington 57 0.5103 0.5073 Kansas U 55 0.5275 0.5050 Dayton 33 0.4733 0.4987 Central Michigan 48 0.4687 0.4670 Northeastern 52 0.4931 0.4508
I am disappointed that Notre Dame and North Carolina aren't playing each other this year in regular season. I think I've missed only one game of this rivalry since 1995 in either South Bend, Chapel Hill or College Cups.
I'm really appreciating and enjoying all the work you guys are doing on this thread. The anticipation generated in this thread will help me get through the next 60 days of no UNC soccer (the World Cup will help as well ). Just out of idle curiosity, do your methods allow you to associate a standard error with any of these calculations? If so, I'd be very interested in how wide the SE might be? It looks like a number of schools are clustered fairly tightly between .6 and .55. Does your model say anything about whether these are statistically meaningful differences? Again, thanks for doing all this work (especially as it makes Stanford and UNC look so aggressive in their match selections).
For individual games, if the ARPI rating difference between teams is 0.01 or less, then the higher rated team wins 46.9% of the time, ties 16.8% of the time, and loses 36.4% of the time, based on 2009 results. These are pretty close to what the numbers would be based on 2008 results, so they appear to be in the right range. When the ARPI rating difference gets to 0.05 or less but more than 0.04, then the higher rated team wins 60.3% of the time, ties 13.2%, and loses 26.4%, based on 2009 results. That may give you some picture of how to evaluate the differences in average opponents' strengths. I'd say a difference of 0.01 doesn't mean much. A difference of 0.05 is significant but not massive.
I now have published 2010 schedules for all of the top 60 teams as rated by the 2009 RPI. That lets me do calculations for each team of (1) the team's 2010 average opponent's strength and (2) the team's 2010 average non-conference opponent's strength. To measure average opponent's strength, I took the average of a team's 2010 opponents' 2009 RPI ratings; and to measure average non-conference opponents' strength I took the average of a team's 2010 non-conference opponents' 2009 RPI ratings. Obviously, teams change from year to year, so one has to bear that in mind, but for a pre-season look at teams' strengths of schedule, this is probably about as good a numerical comparison as one can get. Here are my results, in two formats. The first format puts the teams in order of their 2010 average opponent's strength. The second format puts the teams in order of their 2010 average non-conference opponent's strength. Comparing the two formats can provide some insight into the importance of conference average strength of schedule in relation to a team's overall strength of schedule. The data are in the following order: 2010 strength of schedule rank (based on overall schedule for the first format and non-conference schedule for the second), team, team's 2009 final RPI rank, team's 2010 strength of overall schedule, and team's 2010 strength of non-conference schedule: Ranked by Strength of Overall Schedule 1 NorthCarolinaU 1 0.6063 0.6030 2 Stanford 2 0.6063 0.6023 3 ArizonaState 32 0.5981 0.5747 4 Duke 43 0.5969 0.5616 5 UCLA 3 0.5959 0.5779 6 FloridaState 6 0.5941 0.5648 7 WakeForest 8 0.5933 0.5626 8 SantaClara 12 0.5925 0.6095 9 BostonCollege 7 0.5918 0.5565 10 VirginiaU 24 0.5901 0.5461 11 VirginiaTech 10 0.5893 0.5531 12 OregonU 58 0.5864 0.5473 13 WashingtonState 18 0.5863 0.5531 14 MiamiFL 49 0.5837 0.5333 15 CaliforniaU 30 0.5836 0.5420 16 FloridaU 15 0.5833 0.5958 17 TexasA&M 22 0.5774 0.6162 18 WashingtonU 26 0.5763 0.5332 19 PennState 16 0.5761 0.5812 20 SouthernCalifornia 17 0.5756 0.5336 21 PortlandU 4 0.5746 0.5864 22 GeorgiaU 20 0.5743 0.5733 23 MarylandU 14 0.5739 0.5125 24 WisconsinU 31 0.5721 0.5889 25 Auburn 40 0.5706 0.5596 26 Vanderbilt 51 0.5661 0.5468 27 OregonState 19 0.5658 0.5107 28 ConnecticutU 27 0.5644 0.5592 29 MinnesotaU 37 0.5640 0.5519 30 Villanova 45 0.5628 0.5594 31 SanDiegoU 47 0.5625 0.5587 32 MissouriU 44 0.5618 0.5789 33 Purdue 38 0.5616 0.5471 34 OhioState 21 0.5607 0.5471 35 NotreDame 5 0.5604 0.5859 36 SouthCarolinaU 9 0.5589 0.5418 37 Georgetown 56 0.5588 0.5471 38 Marquette 23 0.5573 0.5627 39 LSU 13 0.5568 0.5330 40 Rutgers 39 0.5554 0.5575 41 WestVirginiaU 25 0.5542 0.5550 42 MichiganState 50 0.5541 0.5280 43 StJohns 41 0.5526 0.5243 44 MississippiU 35 0.5520 0.5199 45 SanDiegoState 29 0.5496 0.5702 46 IndianaU 53 0.5489 0.5165 47 OklahomaState 36 0.5461 0.5468 48 UCF 11 0.5456 0.5789 49 BYU 28 0.5350 0.5567 50 Memphis 34 0.5338 0.5450 51 UAB 59 0.5328 0.5390 52 ColoradoCollege 42 0.5286 0.5312 53 KansasU 55 0.5275 0.5050 54 UNCWilmington 57 0.5103 0.5073 55 UNCGreensboro 60 0.5041 0.5646 56 Charlotte 46 0.4918 0.5270 57 Northeastern 52 0.4863 0.4454 58 Samford 54 0.4835 0.5086 59 Dayton 33 0.4733 0.4987 60 CentralMichigan 48 0.4687 0.4670 Ranked by Strength of Non-Conference Schedule 1 TexasA&M 22 0.5774 0.6162 2 SantaClara 12 0.5925 0.6095 3 NorthCarolinaU 1 0.6063 0.6030 4 Stanford 2 0.6063 0.6023 5 FloridaU 15 0.5833 0.5958 6 WisconsinU 31 0.5721 0.5889 7 PortlandU 4 0.5746 0.5864 8 NotreDame 5 0.5604 0.5859 9 PennState 16 0.5761 0.5812 10 UCF 11 0.5456 0.5789 11 MissouriU 44 0.5618 0.5789 12 UCLA 3 0.5959 0.5779 13 ArizonaState 32 0.5981 0.5747 14 GeorgiaU 20 0.5743 0.5733 15 SanDiegoState 29 0.5496 0.5702 16 FloridaState 6 0.5941 0.5648 17 UNCGreensboro 60 0.5041 0.5646 18 Marquette 23 0.5573 0.5627 19 WakeForest 8 0.5933 0.5626 20 Duke 43 0.5969 0.5616 21 Auburn 40 0.5706 0.5596 22 Villanova 45 0.5628 0.5594 23 ConnecticutU 27 0.5644 0.5592 24 SanDiegoU 47 0.5625 0.5587 25 Rutgers 39 0.5554 0.5575 26 BYU 28 0.5350 0.5567 27 BostonCollege 7 0.5918 0.5565 28 WestVirginiaU 25 0.5542 0.5550 29 VirginiaTech 10 0.5893 0.5531 30 WashingtonState 18 0.5863 0.5531 31 MinnesotaU 37 0.5640 0.5519 32 OregonU 58 0.5864 0.5473 33 OhioState 21 0.5607 0.5471 34 Purdue 38 0.5616 0.5471 35 Georgetown 56 0.5588 0.5471 36 Vanderbilt 51 0.5661 0.5468 37 OklahomaState 36 0.5461 0.5468 38 VirginiaU 24 0.5901 0.5461 39 Memphis 34 0.5338 0.5450 40 CaliforniaU 30 0.5836 0.5420 41 SouthCarolinaU 9 0.5589 0.5418 42 UAB 59 0.5328 0.5390 43 SouthernCalifornia 17 0.5756 0.5336 44 MiamiFL 49 0.5837 0.5333 45 WashingtonU 26 0.5763 0.5332 46 LSU 13 0.5568 0.5330 47 ColoradoCollege 42 0.5286 0.5312 48 MichiganState 50 0.5541 0.5280 49 Charlotte 46 0.4918 0.5270 50 StJohns 41 0.5526 0.5243 51 MississippiU 35 0.5520 0.5199 52 IndianaU 53 0.5489 0.5165 53 MarylandU 14 0.5739 0.5125 54 OregonState 19 0.5658 0.5107 55 Samford 54 0.4835 0.5086 56 UNCWilmington 57 0.5103 0.5073 57 KansasU 55 0.5275 0.5050 58 Dayton 33 0.4733 0.4987 59 CentralMichigan 48 0.4687 0.4670 60 Northeastern 52 0.4863 0.4454
As I think about the overall strength of schedule vs non-conference SOS numbers, what I see is this: For non-conference SOS, 18 of the top 25 teams are from outside the ACC and Pac 10. For overall SOS, only 8 of the top 25 are from outside the ACC and Pac 10. Texas A&M drops from #1 in non-conference SOS to #17 in overall SOS; Santa Clara drops from #2 to #7; Florida from #5 to #16; Wisconsin from # 6 to #24; Portland from #7 to #21; Penn State from #9 to #19;Georgia from #14 to #22; and Auburn from #21 to #25. All the other non-ACC, non-Pac 10 teams in the top 25 non-conference SOS drop completely out of the top 25 overall SOS (Notre Dame from #8 to #35, UCF 10 to 48, Missouri 11 to 32, San Diego State 15 to 45, UNC Greensboro 17 to 44, Marquette 18 to 38, Villanova 22 to 30, Connecticut 23 to 28, San Diego 24 to 31, and Rutgers 25 to 40). Looking at it the other way around, every one of the ACC and Pac 10 teams in the top 25 overall SOS improves its position from the top 25 non-conference SOS. This illustrates the virtual impossibility of teams outside the ACC and Pac 10, even teams from the next tier of conferences, matching the ACC and Pac 10 teams in overall strength of schedule.
Note the highlighted portion. Not that there's anything wrong with it, it just seems that if a team's last 9 games include UCLA, USC, Stanford, & Cal, that it will enhance that team's RPI, for better or for worse. Same thing with the ACC, if 4 of a team's last 10 games include North Carolina, FSU, BC, Wake, it can have a dramatic spike in their RPI, not to mention that the level of competition will make them compete even harder. The corollary to your point would be Notre Dame. They play in a relatively weak conference, which would affect their overall RPI, yet they seem to be in a position to compete for one of the four #1 seeds every postseason, regardless of the level of competition in their conference. What I've seen in basketball, and perhaps this can be adapted to soccer, is that teams will play tough non-conference games in the middle of their conference season. The theory is that this high-level non-conference matchup can only help both teams as they grind it out in their own conferences. So you could have Notre Dame play Protland in the middle of October, for example. I also think that such a game later in the season, as opposed to September, can only help both teams on the field in terms of competition, and also help them off the field with their RPI. The argument against such a high-profile matchup late in the season is that it will distract these teams from the task of winning the conference outright. However, when I think about it, while winning the conference may be a worthwhile endeavor, even in the worst of scenarios, even the 2nd-placed teams in the WCC or in the BigEASY will be offered a post-season invite. Just a thought. Overall, nice analysis, cp
some other things to think about: not every team can schedule the half dozen teams you mention. certainly not every year. that possibility probably just went down with PacX and Big10(12) expansion. They will now have fewer open dates. UP rotates through the top 4 PAC10 as much as it can, (played 5 PAC10 teams last season including Cal, and 2 ACC) but since it played USC and UCLA recently two years in a row those are out, and it is only this year that Stanford and Call are Both in. I'm sure UP would play them more, but there are only so many dates. It's even harder with the top out of region teams. each generally only goes West one weekend a year (and vice versa). An additional complication is that those games are scheduled 2,4 or more years aHead. Penn State's game with Portland was arranged at the end of the 2005 season, as was Stanford's (which I heard rumors about in 2006), but It's pretty hard to judge who the top teams are 4+ years out. Basketball plays non conference games during the conference season to appeal to their selection process, which really doesn't use RPI exclusively or even very much compared to soccer. Polls, ratings, and performance late in the season are much more important. Soccer, on the other hand, uses almost exclusively products of the RPI. And co-operation between conferences is pretty hard as it is. What open dates do the PC10 and ACC, and Big 10(12) have in the middle of the conference season? (none. right now, is my guess. I don't think they would risk tired players on the one game - weekend dates) I agree it's a nice analysis. Good job again, CPT.
The other thing I find interesting looking at how the top 60 teams have scheduled is that most of their schedules appear to be consistent with a pretty good understanding of the RPI and the NCAA's Tournament selection and seeding processes. Teams from outside the ACC and Pac 10 know that they need to have stronger non-conference schedules, on average, than teams from the ACC and Pac 10 in order to be able to compete with them in at large selections and seeding. (Also, they are able to schedule non-conference games against ACC and Pac 10 teams, whereas ACC and Pac 10 teams can't -- at least, practically speaking they can't.) On the other hand, all but the top ACC and Pac 10 teams know that they need to have weaker non-conference schedules than the top teams from outside the ACC and Pac 10, in order to have a decent winning percentage and to be sure to meet the "at least 50% wins" requirement to qualify for the Tournament. The main ACC and Pac 10 exceptions are the teams that expect to win those conferences -- UNC and Stanford (and Florida State and UCLA to a much lesser extent), which expect to have good conference winning percentages to combine with their good conference strength of schedule and so are able to have strong non-conference schedules without taking too much risk. The exception to this "scheduling to the RPI and other NCAA criteria" comes when I look at some of the teams at the bottom third of the non-conference strength of schedule list. I'm looking particularly at UAB, Colorado College, Charlotte, Samford, UNC Wilmington, Dayton, Central Michigan, and Northeastern. If those teams want to be viable competitors for at large slots in the NCAA Tournament, they need to have stronger non-conference opponents than they're currently scheduling -- unless, of course, they don't want to take the risks inherent in playing stronger opponents. If they don't want to take those risks, of course, it suggests that they don't really see themselves as viable at large competitors.
Where does Northwestern's (77th in 2009 RPI) schedule fit in your top 60 teams strength of schedule list?
If Northwestern were in the top 60, their overall average strength of schedule would be 16th and their non-conference opponents' average strength of schedule would be 7th.
Sounds like a tough schedule. Is the difference in the ranking of their conference schedule from other Big ten teams the fact they they don't play themselves? I apologize if you've said this already but do home and away factor in?
Thanks for doing the calculation. Their schedule looked surprisingly difficult. I wonder how many other teams outside the top 60 fall within your top strength ratings. I think someone already asked, but are you planning on doing the list for all of the teams (who have released schedules).
Good questions. Home and away do not factor in, as the RPI as used for Division I women's soccer does not factor in game sites. Game site does make a difference, but so far I have not identified an RPI adjustment for game site that improves the correlation between RPI ratings and game locations. That is a project for after the 2010 season. You are correct that how teams fare when you go from non-conference strength of schedule to overall strength of schedule, as compared to how other teams from the same conference do, has to do in large part with the fact that the team does not play itself whereas other teams from the conference do. As an example, the teams from the Pac 10 other than Stanford get a better boost than Stanford when Pac 10 games are included because Stanford has the highest RPI in the conference. Conversely, the teams from the ACC get a lesser boost than Clemson when ACC games are included because Clemson has the lowest RPI in the conference. Other things that affect how teams from the same conference change rankings, in the transition from non-conference to overall, have to do with the exact average non-conference and overall rankings of the teams and the area of the ratings in which they are located. (The ratings are more spread out at the margins and more bunched together in the middle -- a typical "bell curve" type of distribution.)
This set of calculations is completely manual, so I'm not planning on doing all the teams. Any chance your system could do it? It would be interesting to see the whole array.
I added both the "strength of schedule" and the "non-conference strength of schedule" computed using my unadjusted RPI numbers for the 2009 season. I couldn't find an end-of-season 2009 spreadsheet on your website to use your either adjusted or unadjusted end-of-season numbers. My 2009 numbers don't include the NCAA tournament games. I have the following teams which weren't in the top 60 hitting full schedule and non-conference top twenties: Full Schedule: 6. Arizona 12. Northwestern Non-conference: 5. Boston U. 6. Milwaukee 10. Northwestern 15. Utah Some of the positional rankings don't match at all... e.g. my calculation for Wisconsin's NC strength puts them in a very different position... my 22rd compared to your 6th... (maybe you have counted the exhibition against Marquette?). I also added the URPI average and non-conference URPI average to the individual team pages to see the numbers the from which the averages were calculated... obviously very dependent on my (and the schools') schedules being correct.