2009 ACC Strength of Schedule

It’s time once again to look and see how unbalanced conference schedules affected the ACC season. This is our fourth year with a 16-game conference schedule in a 12-team conference. Our previous looks are listed below, but some of the tables no longer look quite as nice as they did when they were first done.

2006 ACC SOS
2007 ACC SOS
2008 ACC SOS

In previous years, I wrestled the results out of a spreadsheet of my own design. I automated the process as much as I could, but there was still a lot of manual manipulation required….which takes a whole lot of time that I just didn’t have this year. Early on this year, I found that cbssports.com was calculating a “Conference SOS” and displaying on each team’s “RPI Breakdown” page.

I couldn’t find any discussion of this calculation, but my guess is that it uses the basic RPI formula for SOS and only uses conference games. My original goal was to compare my results to cbssports SOS numbers and see if there were any big differences. But life sometimes redirects plans, so we are going to go with the SOS calculations supported with a little common sense review.

Luckily, I copied the SOS numbers before the ACCT just in case CBS decided to throw the tournament games into the mix….which is exactly what they did. So without further rambling, here are the results for the ACC regular season:

acc-sos

I added a column with normalized results to make it easier to see how much change there is from one position to the next. I highlighted third through sixth place because the differences between those four schedules are so small that they are hard/impossible to see in the breakdowns to follow.

Here is a master schedule for the season sorted by SOS vertically and conference finish horizontally. We’ll start with this breakdown and see what sorts of things jump out.

2009-acc-bb-schedule

If you look at the column on the far left and the two columns on the far right, we see how playing UNC twice elevated SOS and playing GT and/or UVA twice dragged it down.

If you look at BC and UNC’s schedules, you will see that both schools came within one game of the easiest possible schedule. Move their second game against Duke (WF) to a second game against GT (UVA) and each school would have the easiest possible schedule.

It would be easy to conclude that the worst teams will always have the hardest schedules and the best teams the easiest….because the teams at the two extremes can’t play themselves. However, this conclusion is more of a tendency, than an absolute fact. The first year that we did this comparison, UNC finished second in the conference and tied for the hardest schedule.

Strictly by the numbers, Duke had the fourth strongest schedule. However, a closer look shows that they played the maximum number of games (6) against the other teams that earned a first-round ACCT bye. The other three teams with a first-round bye all played four games against the top teams in the conference. The only way to increase the difficulty of Duke’s schedule is to move their second game against two 7-win teams to a second game against the two 9-win teams.

I am going to add one more table that parses the games played against different groupings from the standings. I liked the way that this table turned out so I decided to include it. However, this table may prove to be more confusing than enlightening…so study it at your own risk.

parsed-bb-schedule

I think from an overall perspective, the parsing of the games into groups tends to support the SOS rankings. From one team to the next, the SOS ranking is not always so clear (like Duke vs VT), but overall the parsing does a pretty good job at supporting the SOS rankings.

CONCLUSIONS

The NCAAT Selection Committee made an obvious change several years ago and they don’t select very many teams with losing conference records, regardless of the strength of that conference. This conclusion can be supported from a subjective review of the selections or from a statistical analysis (ie the Dance Card). All of this means that it is interesting to look to see how conference SOS may have affected the NCAAT/NIT selections.

So did SOS affect UM, VT, BC, and Clemson’s post-season tournament?

Well did the officiating affect the VT/UNC game in the ACCT?

I think that it is safe to say that in both cases, the “outside influences” had AN effect…but were not were not the only thing that contributed to the final result. BC had an easy conference schedule, but they also had wins against UNC, Duke, and FSU. Miami had a much tougher schedule, but SOS doesn’t explain losses to NCSU and GT. On balance, I think that the officiating in the VT/UNC game had a lot more to do with the outcome of that game than this year’s conference schedule had to do with any team’s final destination.

FOUR YEAR SUMMARY

Here is the four year SOS summary with each year normalized against the toughest schedule for that year.

cumulative-bb-sos

By changing the ranking method this year, you can’t use any sort of averaging or totals of the normalized SOS. However, we can still make some observations:

– Miami has ranked in the top half of the conference in each of the four years. (Though they tied for sixth place in 2007.)

– Duke, NCSU, and VT have ranked in the top half of the conference in three of the four years.

– UVA, Clemson, WF, and BC have ranked in the bottom half of the conference in three of the four years.

Assuming that anyone considered conference SOS when the 16 game rotating schedule was being developed, I suspect that the general consensus was that things would even out over time. However, that conclusion ignores over 50 years of history that says that while teams do move up and down through the standings, there are always teams that are consistently at the top or bottom of the conference for extended periods of time. Thus I expect that the selection of a team’s “primary partners” will have a noticeable effect over time.

An 18-game conference schedule has been discussed before by the conference. As we have said before, this change will probably come to pass….but not until the TV contracts have been properly adjusted. Adding two more games each year should help minimize the difference between the conference’s toughest and easiest schedules.

We’ve had several BB entries recently that should have given everyone an opportunity to vent about the problems in State’s BB program. Keep the griping to those entries and keep the comments here on topic.

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

08-09 Basketball General Stat of the Day

21 Responses to 2009 ACC Strength of Schedule

  1. SaccoV 03/25/2009 at 11:57 AM #

    Interesting look at the numbers, VaWolf. Does the unbalanced schedule involve each team playing 5 home-and-homes and the six other games against teams they’ll only see once? Because I was thinking if the schedule does expand to 18 games, will this open the number of home-and-homes to 7.

  2. VaWolf82 03/25/2009 at 12:13 PM #

    Does the unbalanced schedule involve each team playing 5 home-and-homes and the six other games against teams they’ll only see once?

    Exactly.

    The link in the last paragraph goes to an earlier entry where I looked at a number of different ways to implement an 18-game schedule.

  3. choppack1 03/25/2009 at 12:42 PM #

    An 18 game schedule would go a long way to making the schedule more balanced – allowing you to play 7 teams twice (4 teams once.)

    As it now, you are less likely to face a team twice. (You play 5 teams twice, 6 teams once.)

    Plus if you went w/ an 18 game schedule, you could drop the “partner”.

  4. Classof89 03/25/2009 at 1:04 PM #

    The 18 game schedule split 7-2-2 (i.e. 7 home and homes, 2 home only, 2 away only), with 3 partners instead of 2 is clearly the common sense way to go (since that would put your other eight “nontraditional” opponents into a 4-2-2 split, which is easy to schedule). Of course, if we got Wake, UNC, and Duke as our three partners, then our schedule would always be the most difficult in the conference. ACC would, therefore, be unlikely to put both Duke and UNC together on anyone’s traditional partner rotation.

  5. waxhaw 03/25/2009 at 1:15 PM #

    Great data.

    I’ll toss two more points on the obvious pile:

    1) Automatically playing the best team twice will increase your SOS.
    2) When you are bad, not playing yourself, will increase your SOS.

    Couple of other observations:

    1) Herb’s 10 win ACC season was by far our easiest SOS in a while.
    2) Sidney hasn’t caught any schedule breaks yet.
    3) Sidney and Herb are averaging EXACTLY the same record versus NCAA participants. (which is 4-9)

    Hopefully we will see some improvement next year and an easier ACC schedule would be a great start.

  6. VaWolf82 03/25/2009 at 1:22 PM #

    with 3 partners instead of 2 is clearly the common sense way to go

    I doubt that the ACC will go this way because of the difficulty in getting agreements on who your third partner will be. We’ve been through several ways to split up the 18 game schedule here. It might be worth everyone’s time to read through that discussion again:

    http://www.statefansnation.com/index.php/archives/2008/05/09/more-thoughts-on-an-18-game-acc-schedule/

  7. VaWolf82 03/25/2009 at 1:41 PM #

    On the previous entry, I combined next year’s schedule with Giglio’s predicted order of finish. How much that is worth is obviously a function of Giglio’s accuracy. But if he is correct, then next year will probably be State’s easiest schedule since going to 12 teams.

  8. ryebread 03/25/2009 at 2:11 PM #

    I see three fixes to this:

    1) Double round robin returns: That’s 22 league games though, so I don’t see it happening any time soon. The coaches would much rather beat up on cupcakes. This is viable I think only if BC exits the league and we move to an 11 conference league. In my dream world, this is what would happen.

    2) Split the league in basketball like it is split in football, run the conference tournament like the SEC does: That’s 10 games (home and home) for the teams on this same side and 6 games for teams on the other side (split 3 home, 3 away for each team). The scheduling here is easiest, the schedule stays at 16 league games per team and we really create divisional rivalries (between football and basketball). The unbalanced schedule still remains, but at least it is somewhat fair. As much UNC/Duke love being in the same side in football, they’d probably block this going into place in basketball.

    3) Just add two more games: This doesn’t fix the problem and makes the coaches mad. It protects UNC and Duke though. As such, this is probably the option that the ACC will pick.

  9. 61Packer 03/25/2009 at 3:56 PM #

    I’m in agreement that adding two games to the basketball conference schedule would be good, but before they do that, they need to go back to allowing the Big Four teams to play each other twice each season in basketball and once every season in football. You don’t enhance conference rivalries by killing your oldest and best ones.

  10. choppack1 03/25/2009 at 4:09 PM #

    “This is viable I think only if BC exits the league and we move to an 11 conference league. In my dream world, this is what would happen.”

    Obviously, you’ve never made the trip to Boston. From a college football perspective – or I suspect from a b’ball persepective – the trip is nothing special. However, it’s a great city to visit and well worth the $600.00 or so dollars a trip their costs.

  11. Texpack 03/25/2009 at 4:14 PM #

    An 18 game conference schedule would be a step in the right direction. I don’t like the three partner proposal. I’d prefer a system where you always played at least two of your three partners twice, but you weren’t locked in to playing all three twice each year. I don’t mind playing the other Big Four schools twice each year, to me that is what ACC basketball is all about. What I do mind is the disadvantage that puts us at versus teams that get relatively weak partner schools to play each year.

  12. RagingWolf 03/25/2009 at 4:54 PM #

    Sorry if this has been posted before AND for it being off topic but if you have high definition cable in Raleigh you can see a college football classic between UNC and NCSU in 1986 (free on demand). Go to channel 550. It is condensed down to 50 minutes of highlights.

  13. BSIE80 03/25/2009 at 5:39 PM #

    Good analysis. NCSU degree must be paying off. Statistics or Engineering major??

    I think splitting the teams into divisions is a good option.
    Instead of 2 divisions of 6 teams each.

    What if we went with 3 divisions(North/Mid/South Atlantic) of 4 teams each. ACCT would have 3 division winners as the bye teams, and one wild card (best record/highest ranked team).

    Each division will play home/away game versus 3 other teams in the division = 6 games. Play one game against every team= 8 games. Go to 14 games versus 16 or add 2 randomly selected/rotated schedules.

    North Atlantic
    BC
    UMD
    VT
    Va
    Mid Atlantic
    UNC
    Duke
    NCSU
    WF
    South Atlantic
    GT
    FS
    UM
    Clemson

    I know the Mid Atlantic would be the tough division, but what the heck. The focus should be competition, logic, saving cost of travel, etc…..

  14. Ed89 03/25/2009 at 7:37 PM #

    ^^I’ll toss two more points on the obvious pile:

    1) Automatically playing the best team twice will increase your SOS.
    2) When you are bad, not playing yourself, will increase your SOS.

    True for us. And the opposite is true for Carolina. 🙁

    BSIE80, I’ve thought of the same idea (3 divs), but I doubt it would ever happen. I would especially like it in Football. We would dominate. I think we could call them the North, South, and Tobacco Road divisions. 🙂

  15. howlie 03/25/2009 at 8:52 PM #

    This seems like something a statistician would research once there is absolutely nothing left to research…

    Not meaning to demean the one who did a tremendous amount of work… it’s just there’s no smoke to necessitate searching for a gun here.

    Let’s go surfing instead…

  16. buttPACKer 03/25/2009 at 9:23 PM #

    ed89,

    i never thought about it, but the three divisions is brilliant.

  17. Wolf74 03/26/2009 at 6:15 AM #

    18 game schedule!! Play everyone in your division twice and everyone in the other division once. Seed each division individually for the ACC tournament based on regular season. Then each division plays it out in the tournament and the winners out of each Division play each other in the ACC Championship Game.

    If you cann’t play each team twice, that is the way to do it.

    Unfortunately, Duke and UNC (who just adore each other) would block this because it would eliminate them playing in the championship. However, for anyone to ever embrace the division concept, it must be done. Doing it one way in football and another in basketball is just plain crazy (or sucking up to the color blue).

  18. waxhaw 03/26/2009 at 6:59 AM #

    Howlie — I don’t think the point of the research is to say that there is a conspiracy to give us a tough schedule. As I pointed, we are going to have a slightly tougher schedule by nature due to our partner being UNC-CH and us not being very good.

    I do think it is interested data to see the true performance of a team. In the old round robin, a team with 9 wins was better than a team with 7 or 8 due to playing the same teams. In the new system, it’s possible a 7 or 8 win team could have beaten better teams and a 9 win team got there by playing the bottom third of the conference twice.

    (Of course, injuries, teams being hot at different times, home/away breakdown, etc all affect this as well.)

  19. VaWolf82 03/26/2009 at 7:24 AM #

    there’s no smoke to necessitate searching for a gun here.

    How about the year where UVa tied for the regular season championship and also had the easiest conference schedule? The point is, you don’t know if you don’t look.

  20. Daily Update 03/26/2009 at 7:39 AM #

    I have never really analyzed divisions, but I don’t like permanent partners (unless we can choose the weakest basketball programs in the league).

    When the ACC expanded, they threw tradition out the window. In the new ACC and with our basketball tradition killed by our administration, we have to support ideas that help us win games. Playing Duke, UNC and Wake twice every year as an official conference game would be a huge handicap for our basketball program.

    All scheduling in the new ACC should be based on an equal rotation. Nobody should be partnered with any program because those partnerships are advantages to some programs (Wake in football gets Duke) and disadvantages to other programs (NC State and Maryland play Duke and UNC more than any other team in the conference in basketball).

    Nothing irks me more than when someone makes the comment that our job is the toughest in America because of Duke and UNC. It is a ridiculous argument not supported by any facts except for our “partnership” with UNC in basketball that our fan base supports.

    We heard Lee Fowler discussing scheduling Duke in football as an OOC game just last year. It might be the most brilliant idea the man has ever come up with. Unfortunately, I seriously doubt Duke will agree to do this with the current state of their program. Duke knows that they need to schedule wins and that generally speaking they aren’t going to beat us very often. So right now, it isn’t in the best interest of their program to schedule us as an OOC game, but at least that option is there in the future.

    What I want are annual home and homes with Duke and UNC to be options in years where the ACC rotation doesn’t have us playing them twice as conference games. If our team is looking particularly strong or the OOC schedule is particularly weak, then put the extra game against Duke and/or UNC on the schedule. Because we lose 80%+ of our games to do Duke and UNC, then these games won’t hurt us in the ACC standings but will still help our overall SOS and be fun for our fans (Am I the only one that doesn’t enjoy playing these two teams?) IMO, we would be better positioned to make the NCAA tournament with these extra Duke/UNC games as OOC games which gives us opportunities to make runs in March which is what will give our program a chance to establish itself.

    There is no way NC State should support a partnership with UNC that mandates two games per year that count in the conference standings. It isn’t in the best interests of our program.

  21. Alpha Wolf 03/26/2009 at 3:01 PM #

    Haven’t read the comments, but Joe Giglio featured this article on ACC Now. Well done, VA Wolf!

    http://blogs.newsobserver.com/accnow/sorting-through-the-schedules#comments

Leave a Reply