Fired Bill Carmody: More RPI talk

Friday, January 22, 2010

More RPI talk

(if you don't care about RPI and just want to see Kevin Coble hit a backwards shot from 3/4 court on one leg, click here)

Sippin' on Purple took a detailed look at NU's bubble status, so i won't bother trying to give my own interpretation of the same numbers when you can just read it here.

but i do want to spend some more time on Northwestern's strength of remaining schedule and RPI. every March people get all pissed off when the RPI is mentioned, as if some mysterious formula is replacing the good ol' eyeball test. look, RPI is not the be-all, end-all for determining tourney bids and seeding. but the math behind it does make sense - who did you play? and who did you beat?

i don't think we can overstate how bleak the situation is for our Wildcats. it's not just that NU only has five top 100 opponents left. it's that the most of the rest are really REALLY bad. check out the current RPI numbers for the remaining opponents:

77 Illinois
55 @ Minnesota
21 @ Michigan St
117 Michigan
192 Indiana
178 @ Iowa
55 Minnesota
233 Penn State
11 @ Wisconsin
178 Iowa
233 @ Penn State
341 Chicago St
192 @ Indiana

assuming those numbers stay about the same, a win(s) over Penn State would come in behind North Florida and Northern Illinois. and Chicago State at 341 is only six spots ahead of the bottom of Division I and is four spots BEHIND the 0-19 Bryant Bulldogs. NU's RPI has been right around 60 for a few weeks and there simply won't be much opportunity to improve that.

using those RPI numbers as unofficial rankings, let's assume NU beats the worse teams and loses to the better teams. that leads them into the conference tournament at 22-9 / 10-8. as SoP explained, that's a good record on paper, but it falls apart once you look at who they would have beaten. and sure, Northwestern *could* win one of those games against MSU, Minnesota or Wisconsin; but they also sure as hell *could* lose one or more against Indiana, Illinois, Iowa, Michigan or Penn State. ok, maybe not Penn State.

finally, here is some RPI history about at-large bids (ALB) and why that 22-9 record and ~60 RPI probably wouldn't be good enough:

The highest rated team left out: #21 Missouri St, #30 Hofstra (2006), #30 Air Force (2007).
... among BCS conferences: #40 Cincinnati (2006).

The lowest rated teams to get at-large bids: #62 Arizona (2009), #62 Iowa St and #63 NC State (2005), #63 Stanford (2007).

Fewest wins to get an ALB: 17 (Alabama - 2006).

Most losses to get an ALB: 14 (Arizona - 2008).

Worst record to get an ALB: 18-14 .563 (Arizona - 2008)

Number of times a team with at least 20 wins has been left out: 109
... among BCS conferences: 22

Most wins left out: 26 (Robert Morris - 2008, Niagara and Creighton - 2009)
... among BCS conferences: 23 (Florida - 2009)

Best record left out: 22-5 .818 (Stephen F. Austin - 2008)
... among BCS conferences: 21-9 .700 (West Virginia - 2007, South Carolina - 2009)


Anonymous Brendon said...

Well-considered post. Tough to swallow for the pie-eyed NU fans out there, but many of us have been indicating as much for weeks. The Big Ten's poor non-conference play is hurting Illinois, Minnesota and especially Northwestern.

10 more wins for NU to be strongly considered.

1/22/10, 10:24 AM  
Blogger #1 blogger said...

i should have mentioned, all of the RPI info came from Jerry Palm's site-

membership is well worth the $25 a year

1/22/10, 11:12 AM  
Blogger Loretta8 said...

was just about to ask where you got all that RPI history, as I spent a long time searching but couldn't find anything like that.


1/22/10, 11:22 AM  
Blogger Steve Kornacki said...

For what it's worth, at least two at-large teams have had worse records than '08 Arizona: 1991 Villanova and 2001 Georgia were both 16-14. Anyway, good site, hope your team makes the tournament this year....

1/23/10, 9:55 PM  
Blogger #1 blogger said...

@ Steve Kornacki - the stats i referenced are from 2005-2009. the RPI formula changed significantly in 2004 so information prior to that change is mostly irrelevant.

1/25/10, 9:23 AM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home