Conference SOS

As soon as ACC decided to go to an unbalanced basketball schedule, it was obvious that some teams would get “harder�? schedules and some would get “easier�? ones. So now the time has come to try and figure out which is which. I decided to compile the ACC wins by each team’s opponents and use that to rank the conference schedules. So subject to potential fat-finger mistakes, here are the teams sorted from hardest to easiest conference schedule:

ACC

Opponents

SOS

TEAM

W

L

Wins

1

Virginia Tech

4

12

140

2T

Miami FL

7

9

139

2T

Wake Forest

3

13

139

4

Maryland

8

8

132

5

Georgia Tech

4

12

131

6

North Carolina

12

4

130

7

Virginia

7

9

129

8

Florida St.

9

7

126

9

NC State

10

6

123

10

Duke

14

2

118

11

Clemson

7

9

116

12

Boston College

11

5

113

If you want to double-check anything in this entry, the team names link to their schedule and results at kenpom.com. Trying to compare Duke’s SOS to WF’s is a complete waste of time and meaningless as well. Duke get’s “penalized�? because they can’t play themselves….and WF gets the “benefit�? of not playing the last-place team. So let’s take a closer look at the top and middle of the ACC and see if there is any meaning to using total opponents’ wins to rank SOS: TOP FOUR TEAMS – When you look at the SOS ranking for BC, State, and UNC…you see that UNC had a substantially tougher schedule than either State or BC. This difference is also borne out by breaking down the schedules:

Games Against

Top 4

Bottom 3

ACC SOS

UNC

5

3

#6

State

5

5

#9

Duke

4

5

#10

BC

4

6

#12

– UNC”s scedule is just about as hard as theoretically possible. One of the few ways to make UNC’s schedule harder would be to take one of their games away from one of the 7-win clubs and schedule a second meeting with BC. (UNC had the absolute minimum number of games against the bottom three teams.) – While Duke does get penalized in my ranking system by finishing first….they also clearly got an easier schedule than UNC or State. – BC got to “max out�? with six games against the three worse teams in the league. They also played only one game more than the minimum against the three other teams at the top of the conference. THE MESS IN THE MIDDLE

Games Against

Top 4

Bottom 3

ACC SOS

Miami

7

3

#2

MD

6

4

#4

UVA

5

4

#7

FSU

5

4

#8

Clemson

4

6

#11

– Once again the schedule breakdown gives the exact same order as the over-all ranking. – While Miami and Maryland played tougher schedules than their nearest competition, both teams lost multiple games against the other teams stuck in the middle…and OOC games as well. (For instance, they both lost to Temple.) An easier schedule wouldn’t have had either team in contention for the conference title, but could have made getting into the NCAA tournament easier (but not necessarily a certainty). – While Dave Leitao has clearly done an outstanding job at UVa this year, I think that when you include the conference SOS into the evaluation…..Roy wins COY going away. (The preceding conclusion was written over a week ago…..is there any doubt now?) – How easy was Clemson’s schedule? They were the only team in the conference that played Duke, UNC, State, and BC only one time each. They also played the maximum possible number of games against the worse three teams in the conference. It doesn’t get much better than that. CONCLUSION Much like State’s season to date, there is nothing exciting to say about its schedule. State played five games against the top three seeds in the ACCT (going 1-4) and five games against the bottom three seeds (going 3-2). Thus State’s schedule was neither particularly hard nor easy…it was really pretty balanced. Since the unbalanced schedule isn’t going to disappear, it serves no purpose to develop a long diatribe outlining its flaws. One thing that I don’t like about the unbalanced schedules is that it will be more difficult to detect meaningful trends based on the regular season results. Note that I said difficult, not impossible…who knows what we might come up with in the future? 🙂

About VaWolf82

Engineer living in Central Va. and senior curmudgeon amongst SFN authors One wife, two kids, one dog, four vehicles on insurance, and four phones on cell plan...looking forward to empty nest status. Graduated 1982

General NCS Basketball

28 Responses to Conference SOS

  1. VaWolf82 03/05/2006 at 6:01 PM #

    One word of clarification.

    When I started this about two weeks ago, there was a huge gap between the top 4 teams and the “mess in the middle” . I would have never guessed that FSU would finish within one game of State in the conference standings. So my break between the “top” and “middle” may not be ideal…but you get what you pay for. 😉

  2. BJD95 03/05/2006 at 6:51 PM #

    Great analysis – very interesting. Certainly puts some beef behind my general feeling that BC is at least somewhat fraudulent.

  3. Wxwolf 03/05/2006 at 6:51 PM #

    Interesting analysis, thanks! I think one trend to watch with the unbalanced ACC schedule will be to see how the NCAA selection committee treats 9-7 and 8-8 ACC teams. In the past, those teams have largely gotten in based on the rigourous round-robin conference schedule. I imagine now those types of ACC records won’t provide the “lock” they once did for the NCAAs. Of course, to counter that fact is the extra round of the ACC Tournament for the bottom 8 teams, which provides the chance for an extra victory there.

  4. Jeff 03/05/2006 at 7:39 PM #

    This is FANTASTIC!!! Thanks so much. First time that I have seen this anywhere.

    This will be a key component of future analysis that will need to be considered in the future.

  5. Sammy Kent 03/05/2006 at 11:00 PM #

    And Clemson still finished ninth. Wow.

  6. PACDADDY 03/05/2006 at 11:14 PM #

    Thanks VA….I was waiting for that breakdown. That is good stuff! I was going to do that myself, but it would have taken me until next season. 🙂

  7. william 03/05/2006 at 11:39 PM #

    I am a believer in balanced schedules, but the RPI and Pomeroy ratings generally do indicate who is playing and doing well against good schedules. Carolina has to play Duke and Maryland and State twice each year. That has both benefits and detriments. Essentially, the only benefit that comes from winning against weak competition is a good ACC tournament seed, since the Conference, unlike the NCAA, does not correct for SOS. However, given the monsterous beast that the conference tournament has turned into, it is hard to take the tournament seriously anymore. If Duke is smart, they will take it easy and play some of the guys on their bench and get ready for the real tournament.

  8. class of '74 03/06/2006 at 6:11 AM #

    22 conference games would fix this. And for Coach K and whoever says this would hurt the development of their team I say bohunk! It would limit the number of cupcakes that’s all.

  9. Jeff 03/06/2006 at 8:01 AM #

    I agree with William on emphasizing that computer models succeed by incorporating SOS (and performance vs those SOS) into their rankings.

  10. RickJ 03/06/2006 at 8:56 AM #

    “Carolina has to play Duke and Maryland and State twice each year.” I know Carolina played these 3 teams twice this year but I don’t think they will play Maryland twice every year. It is my understanding that every team has 2 teams they will play twice every year; ie. State has UNC & Wake – UNC has State & Duke and Duke has UNC & Maryland. I could be wrong about this.

  11. VaWolf82 03/06/2006 at 9:06 AM #

    Rick is right….UNC plays Duke and State twice every year.

    The ACC released what it calls a three-year model. I don’t know if that means that the rotation will be complete in three years or if the ACC intends to re-evaluate the master schedule in three years. In any event, here is the link to the conference schedule for ’06-’08”

    http://graphics.fansonly.com/photos/schools/acc/sports/m-baskbl/auto_pdf/model-three-years.pdf

  12. Nick 03/06/2006 at 9:06 AM #

    The ACC could eliminate all discussion of SOS in ACC basketball standings by instituting divisional play. Each team could play home-and-away with teams within its division and one game each against each team in the other division. Each team would play 16 ACC games as now, but the same 16 ACC games each year. A couple of “rivalry” games would suffer as a consequence, but it would be straightfoward, above board, and immune to manipulation or the perception of manipulation, which is not now.

    The ACC tournament could be run as two parallel 6-team tournaments to crown each division champion who would then meet in the ACC Championship game to decide the ACC Champion.

  13. lumberpack 03/06/2006 at 9:33 AM #

    Nick-I like three divisions

    North, Big 4 and South

    6 divisional games and 8 out of divisional games for a 14 game schedual

    How do you make the two divison split? You have to split the Big 4 to do that.

  14. RAW 03/06/2006 at 9:55 AM #

    I like lumberpack’s idea

  15. Jeff 03/06/2006 at 10:41 AM #

    lumberpack’s idea rocks

  16. VaWolf82 03/06/2006 at 11:05 AM #

    immune to manipulation or the perception of manipulation, which is not now.

    I think that you’ve probably overstepped here. The two designated rivalries were mostly well-done and preserve most of the traditional games that the fans of the schools want to see. The ACC then set up a rotation so that over the long run….things will even out. There is no basis for claiming that anyone manipulated the schedule.

    In the short run of a given season….Clemson will get an easy schedule and Miami will get an unusually difficut one….but no system short of a complete round-robin will prevent that from happening.

  17. Charlotte_Eagle 03/06/2006 at 1:51 PM #

    BC fan here. We were discusing this analysis and someone on our board had the suggestion of subtracting out the # of losses that a team was directly responsible for.

    Without it, top teams are always going to have a lower stregth of schedule than the bottom teams becuase they lost fewer games within the conference. A proposed revised ranking would look like this:

    Miami 130
    VT 128
    UNC 126
    Wake 126
    MD 124
    UVA 120
    FSU 119
    GT 119
    NC St 117
    Duke 116
    BC 108
    Clem 107

    Credit to this sugggestion goes to Orphio from our rivals free board.

  18. VaWolf82 03/06/2006 at 3:26 PM #

    Glad to know that fans from other schools check us out from time to time. Thanks for taking the time to post your results.

    IIUC, I looked at something like this, but gave up. I took two teams….State and VT and calculated an adjustment. I got 128 for VT (same as your board), but got 112 for State (your board got 117). I decided against this technique because it required so much manual manipulation of the data….and the teams at the top are still penalized because they can’t play themselves.

    Probably the best way to determine SOS is to use the exact same formula used by the RPI formula….which is 2/3* Opponent’s winning percentage (with games against the team in question removed) + 1/3 * Opponents’ Opponents winning percentage. I have thought about dropping an e-mail to kenpom.com and asking if he has ever considered doing this for conferences that play an unbalanced schedule.

    Let’s face it….the only time conference SOS really matters is for bubble teams. What would people be saying about Miami if the second loss to State and BC were magically turned into wins against GT and WF? Probably the exact same things that are being said about FSU.

  19. VaWolf82 03/06/2006 at 3:54 PM #

    After taking a second look, the BC fan did something different than I looked at doing.

    They took the total number of opponent’s wins and then subtracted the losses from each team. For example, Duke’s opponents had 118 wins and Duke had two losses. The BC fan took 118-2 = 116 and then sorted on this number.

    But Duke played UNC and FSU twice. What I looked at doing would have subtracted “4” from Duke’s total because the overall wins of UNC and FSU would each have been reduced by one….and Duke played each team twice.

  20. Charlotte_Eagle 03/06/2006 at 4:07 PM #

    Got tuned in to your site by eagle in atlanta’s blog. Given our newness to ACC hoops your historical data driven insight into the ACC and which teams make the tourney has been a very interesting read. Keep up the good work.

  21. Nick 03/07/2006 at 8:14 AM #

    VaWolf82 is right. I overstepped with “immune to manipulation or the perception of manipulation”. Probably should have said “immune to perception of unfairness”. Why analyze SOS if not to justify/debunk the suspicion of unfairness? The current system insures that we will have SOS discussion every year, because every team plays a different schedule and then we rank them 1 to 12 and seed the ACC tournament based on this ranking.

    If the ACC had two divisions, each team within the division would play the same schedule, home-and-away games with teams in the division and one game each with the teams from the other division. Every year the same schedule. And over time, divisional championships would come to mean something. Banners would go up. And rivalries would develop.

    I would keep the current divisions and swap Maryland and Carolina to break up Duke and Carolina. Another option is to sway State and Carolina.
    Either way it wouldn’t affect the football much.

    The trouble with the three division format is how would you handle the ACC Tournament? Three division champions and a wild card? How would you decide the wild card? Most wins in head-to-head games among the divisional runners-up? That could work, I guess.

  22. VaWolf82 03/07/2006 at 8:46 AM #

    Why analyze SOS if not to justify/debunk the suspicion of unfairness?

    The analysis is just an attempt to see what the “truth” is. For example, we see that UNC played a hard schedule and to finish second against that schedule is a pretty darn good accomplishment. Likewise, we see that Clemson’s best record in a good while came against the easiest possible schedule….thus not so impressive.

    The current system is pretty “fair”, because with the exception of two partners, the rest of the teams rotate from year to year. As I said in the opening paragraph, it’s pretty obvious that an unbalanced schedule will produce “easier” and “harder” schedules.

    A divisional format doesn’t eliminate the potential for “unfair” schedules. There would still be times where the second or third best team in one division would be better than any team in the other division.

    If you are concerned about being fair….a complete round-robin is the only solution.

  23. Nick 03/07/2006 at 10:56 AM #

    A complete round robin in not going to happen. I think a system with two divisions would be better than the current system. Everyone within the division plays the same schedule. That’s important if you are going to have standings that mean objectively meaningful. That was all I was saying. I will not beat this dead horse further.

  24. VaWolf82 03/07/2006 at 11:48 AM #

    A complete round robin in not going to happen.

    I agree.

    I think a system with two divisions would be better than the current system. That’s important if you are going to have standings that [are] objectively meaningful.

    I’m not sure if I agree or not….but let’s assume that you are right. Do you think anyone cares about conference or divisional standings?

  25. Nick 03/08/2006 at 10:10 AM #

    Yes, I think people care about conference standings. It’s how people judge how well the different teams performed during the year. It’s how the ACC decides the seedings for its tournament.

    I just think that standings are more meaningful when the teams play the same schedule. Since the teams are unwilling/unable to play the same schedule now that the conference expanded to 12 teams, the best alternative is to divide the 12 teams into two divisions within which the teams play the same schedule year in and year out. Over time, I think this would foster competition. Divisional championships will come to mean something over time and divisional loyalties will develop. Also, I think it would foster rivalries where they might not now exist.

Leave a Reply