An RPI article that even VaWolf can appreciate

General13

I ran across this work from “Roll Bama Roll” and thought the community would enjoy it. (For more general college basketball discussion from this weekend I recommend that you click here.)

RPI advantages BCS Schools? Hardly

Some in the national media have made off-handed comments about the RPI advantaging BCS-conference schools, saying fans of schools like Virginia Tech, Alabama and Nebraska are only complaining about the RPI because their teams happen to be ranked lower in the RPI than other NCAA Tournament hopefuls.

Not only are these “experts” wrong, but if you believe that the Pomeroy ratings are a much better tool at gauging teams–as most of them profess to do–then ironically the opposite is actually true. The RPI ratings seem to favor teams that aren’t from BCS conferences, relative to more sophisticated tools like Pomeroy. Now of course, Pomeroy is itself not a perfect tool, but it is widely considered to be a much stronger measure, and in fact is often relied upon heavily by many oddsmakers. For these purposes we’ll simply use the Pomeroy ratings as a comparison tool.

The following data is based on the Pomeroy ratings of 2/26 per kenpom.com and the RPI ratings of 2/26 per warrennolan.com. The only teams examined are those ranked in the top 100 of the RPI, since those are the only teams for whom the RPI could potentially be relevant when it comes time for the NCAA and NIT selections.

The full data is below the jump, but here are the aggregate results:

About StateFans

'StateFansNation' is the shared profile used by any/all of the dozen or so authors that contribute to the blog. You may not always agree with us, but you will have little doubt about where we stand on most issues. Please follow us on Twitter and FaceBook

General

13 Responses to An RPI article that even VaWolf can appreciate

  1. choppack1 02/27/2011 at 9:44 PM #

    I’d actually go as far to say that the RPI does exactly what it is supposed to do – gives you more non-BCS teams in the field of 64. Look, it’s not an accident that this calculation is used…And let’s look at how f’d up the RPI is.

    According to the RPI, if Team A beats Team C 100-0, and Team B beats Team C 50-49 – team b and Team A are equal.

    I simply don’t understand how over the course of 25 for the 150 or 200 eligible for the NCAA tournament at the beginning of the year, you can throw it such valuable data, unless of course, the data stops you from accomplishing what you want to accomplish.

  2. VaWolf82 02/27/2011 at 10:26 PM #

    So much for not calling out posters. 😉

  3. VaWolf82 02/27/2011 at 10:34 PM #

    I don’t have time to digest all of that right now. But the last change in the RPI calcs definitely elevated teams at the top of mid-majors relative to teams in the middle of power conferences. This change is even easy to explain without numbers or tables.

    The last change gave more credit to road victories than to home wins. So how does that affect things? Here are two givens:
    – Teams at the top of any conference are going to get a lot of road wins during conference play.
    – Teams in the middle of any conference tend to win more at home and less on the road.

    Thus the last change in the RPI formula tended to elevate top teams in mid-majors and depress the rankings for teams in the middle of the conference. Strangely enough, the professors behind the Dance Card have detected no change in the Selection Committee’s decision making process. In fact, they still use the old RPI results that are supplied to them by Jerry Palm at collegerpi.com.

    I don’t know how or even if that ties into Jeff’s article. I may be able to get to it tomorrow.

  4. choppack1 02/27/2011 at 10:38 PM #

    VaWolf – One thing I don’t understand – if I’m doing schedule for NC State or VaTech- I’m playing as many UNC-G’s or Elon’s in the GSO Coliseum as I can. I’m playing UNC-A in Boone. I’m playing Wofford in Greenville…I’m playing the Citadel in Charleston.

  5. VaWolf82 02/27/2011 at 11:19 PM #

    During one of the many OOC scheduling discussions we’ve had around here, someone counted up and State played something like 10 OOC home games every year. Consistently like that usually denotes a plan…ie the budget plan for the athletic dept. So I doubt that moving cupcakes games to the Greensboro Colisieum (neutral court) or playing away (like Fowler’s plan for ECU) is going to become part of the master plan for DY.

    I don’t have a program that would let me play around with numerous options, but I did the numbers by hand for ECU several years ago. Depending on where you were ranked, beating ECU in G’ville vs beating them in Raleigh could have made no difference up to a couple of spots.

    If you want to help your OOC schedule…just avoid scheduling horrible programs. Some programs are up and down, some are usually up, and some are nearly always down. NEVER, NEVER, EVER schedule games against bad programs no matter what conference they are in. Eliminate the RPI anchors and you’ll do more to help your RPI than trying to play games with neutral/away OOC games.

  6. FunPack 02/28/2011 at 12:23 AM #

    Good find. Interesting stuff. Wasn’t it Billy Packer who criticized the NCAA selection committee a few years back for their selection of several mid-majors over middle-of-the-pack big-conference teams? Looks like he may have had a valid point.

  7. LRM 02/28/2011 at 8:32 AM #

    Packer often had valid points, he was just too pompous in his delivery.

    When I look at the RPI, all I see is how high State could have been rated, with a couple wins over several of the Top 20 RPIs we played OOC (Georgetown, Wisconsin, Syracuse, Arizona), plus Duke and Carolina twice.

  8. VaWolf82 02/28/2011 at 9:08 AM #

    Winning three games against UMD, FSU, and VT would give State the same RPI ranking as winning three games against Duke, UNC, and Georgetown.

  9. LRM 02/28/2011 at 9:12 AM #

    I see. So, basically, it’s even worse than it looks, considering none of those are even in the Top 50 RPI.

  10. haze 02/28/2011 at 10:23 AM #

    In the end, if you need to be counting on the RPI or KenPom to justify your appearance in the tourney, you’re (almost by definition) not good enough for your tourney appearance to lead to anything meaningful. Thus, DY should certainly stay with OOC home games ($) and we should also schedule decent OOC games (more $). There is a reason why Va Tech, Bama and NC State lead the world in analysis of NCAA invites… and it’s a sickly reason.

    The cool part of this analysis is the mid-major debate. I whole-heartedly agree with rewarding OOC road wins and I think it’s perfectly fair to reward mid-majors that travel successfully in Nov/Dec. However, VaWolf’s point regarding in-conf road win bias is telling and suggests an equal weighting for in-conf games is probably better.

  11. VaWolf82 02/28/2011 at 11:09 AM #

    Ok, I’ve had time to digest the article Jeff linked. I see several problems.

    – The assumption is made that kenpom is a better system in evaluating teams than RPI. This may be true, but it seems like this assumption needs to be proven before you can use it to base ALL of the other analyses on.

    – Even if kenpom is proven to be a better system for ranking teams nationally, so what? The RPI is one of the things that is used in selecting and seeding the teams in the NCAAT. It is not the only thing used. If you want a point worth discussing, then determine if the NCAAT SELECTION AND SEEDING PROCESS is biased, not just some small piece of the bigger puzzle.

    During the Herb years at State, we learned a lot about the NCAAT selection process:
    – Big wins against Top 25 and Top 50 wins are required to get into the NCAAT – but it doesn’t really take that many.
    – Big wins against Top 25 teams in the conference tournament pays really big dividends.
    – Big OOC wins are not required for teams in the power conferences.
    – Bad losses do not doom bubble teams to the NIT.

    If anything, playing in a power conference makes it easier to get into the NCAAT because you have so many more chances to get the big wins that you need. In a lot of cases, mid-majors are going to have to go on the road in their OOC schedule to get those big wins. They are also going to have to correctly guess which teams from the power conferences that they need to play.

  12. wufpup76 02/28/2011 at 3:41 PM #

    “If anything, playing in a power conference makes it easier to get into the NCAAT because you have so many more chances to get the big wins that you need. In a lot of cases, mid-majors are going to have to go on the road in their OOC schedule to get those big wins. They are also going to have to correctly guess which teams from the power conferences that they need to play.”

    ^Bonita … That sums it all up right there. I will almost never complain about a middie having a good rpi. Beyond actual results, I still think the ‘eye test’ is the best way to determine a team’s validity. The rpi is a reasonable tool though for quantifying a lot of subjective data.

  13. Wolf74 02/28/2011 at 8:33 PM #

    “I’m playing UNC-A in Boone.”

    I’m not so sure App State would go along with that. Might want to play UNC-A in their home town of Asheville. In fact they have a very nice new coliseum that is opening next season. Guess who the opening game is – UNC-CHeats. Our own Ed Beidenbach is the head coach. I have talked to him several times over the last few years and he is still a big NCSU fan and keeps up with the games and stays in contact with the coaches.

Leave a Reply