RPI, BPI, Kenpom....What's It All Mean???

Great game between WVU and Kansas last night.  WVU had them, let them back in and finally lost in OT in Allen Field House in Kansas…Too bad for Huggins as that would have been a possible seed changing victory!!!

I was looking at Michigan’s Kenpom rank on Monday (Which is here)

 

And Stumbled upon this (This is AWESOME!!!):

 

So, if last year’s 15-Seeded MTSU and 2-Seeded Sparty were to play again, Sparty is a DOG!!! 

HAHAHAHAHEEEHEEEEEEE!!!!    

Which got me to thinking about and doing a little research on Kenpom, BPI and RPI in an attempt to clear up some of the confusion surrounding all of these statistical measuring systems. 

Michigan’s RPI sits at 61 as of this post.  But Michigan is RANKED somewhere in the 30’s per the coaches poll.    Michigan’s BPI (ESPN’s answer to RPI that was created in part by Duke Alum and ESPN commentator Jay Bilas) is 24, the spread being 37.  And their Kenpom at 31 is in line with their BPI and coaches ranking….

So why the discrepancy?  For one thing, RPI is really old and outdated having been introduced in 1981.  It doesn’t use any predictive analysis, scoring percentage or coach’s opinion and focuses on what’s occurred up to this point in time.  It more or less just looks at wins and losses.  It doesn’t adjust for points scored in garbage minutes, scoring average per possession, or points scored against better defenses and other things that can be done very quickly today with a simple Excel spreadsheet. 

Which begs the question, “Why in the hell does the Selection Committee use the RPI instead of the BPI or Kenpom?”  

Well, Kenpom and BPI are more predictive in nature and RPI is, and (I use this term VERY loosely), “a solid look at what has happened.”  Your RPI can move as you beat better teams, but until you do, that’s where you’re stuck.  Michigan has a 37 point discrepancy.  Colorado is the worst with a discrepancy of around 60 points…  When you see discrepancies like that, it usually means an SOS or some other oddity is holding a low BPI/Kenpom ranked team back at a higher rank in the RPI (lower being better).  So they may be playing a large part of their SOS late in the season (Michigan is).  In fact, Wiscy, Purdue and Northwestern (barely) rank ahead of Michigan in Kenpom with Minnesota, Rutgers and Nebraska below.  Meaning that Michigan should, in Vegas terms, go 3-3 in the next 6 games.  Therefore, beating Wiscy would be a huge upset Thursday night.  I really wish Wiscy hadn’t lost to Northwestern Sunday...However, 4 out of the 6 teams are ahead of Michigan in RPI, so beating 2 of the lower RPI teams should do wonders…

Again, why doesn’t the committee use the other statistically more accurate systems?  They do, but they’re just not allowed to “talk about using them.”  Often, when a team that looks deserving via RPI is left out of the tourney, check their Kenpom or BPI metric.  If the Kenpom/BPI are the same or higher than RPI, chances are the team isn’t getting in no matter what they’ve accomplished (assuming they’re a bubble team to begin with).

But when a high RPI sneaks into the tourney, check their Kenpom or their BPI.  Chances are there’s a large “positive” discrepancy for that particular team that “Defied the odds.” 

Where it gets really interesting is, for example, if the committee had to seed a 4^th and final SEC team with identical 14-11 records (as of today).   Georgia has an RPI that is 10 points better than Tennessee’s.  But Tennessee’s Kenpom position/predictor is 10 notches higher than Georgia’s!  So who would get in?  Tennessee lost the only H-T-H matchup this season 76-75 on their home court…

That’s why we have a committee to look at big wins, etc….

Ain’t March Madness Great!!!

 

Opponent     Location       Date   Day and Time          Result

Indiana          Home             1-26    Thursday, 9:15        90-60 Blow-Out

Sparty             Away              1-29    Sunday, 1:00           70-62 Loss

Suckeyes     Home             2-4      Saturday, 6:00         70-66 Loss

Sparty            Home             2-7      Tuesday, 9:00         86-57 BLOW-OUT

Indiana          Away              2-12    Sunday, 1:00             75-63 SWEEP OF IU!!!!

Wiscy             Home             2-16    Thursday, 7:00

Minnesota     Away              2-19    Sunday, 7:00           

Rutgers          Away              2-22    Wednesday, 6:30    

Purdue          Home             2-25    Saturday, 4:00

Northwestern Away             3-1      Wednesday, 7:00

Nebraska       Away              3-5      Sunday, 8:00

If I would have waited about 10 minutes before hitting send on Monday’s e-mail, I would have had Lunardi’s updated “Bracket Watch” that has Michigan in as an 11-Seed.  USA Today has Michigan in as a solid 10.  That’s a HUGE jump from the 2^nd 4 Out (Michigan is the one of the two new teams to the field of 64 as of 2-13), but a blow-out win against Sparty and a Road Victory at Indiana can help a lot.

(Joe Lunardi either has someone on the inside or has an incredible feel for the committee’s tendencies year over year.  I trust his “4 byes, last 4 in, first 4 out and next 4 out analysis”….)

Sparty is still in as a 10-Seed and on the list of 4 byes, but I think that’s on shaky ground, based on the analysis I just gave above.  Sparty has a great RPI at 41 (even after getting blown out by Michigan) but a BPI and Kenpom that are 10 to 20 points WORSE than their RPI.  The only thing holding Sparty up is the Non-con SOS (Where all 4 of Sparty’s Big NON-CON Losses actually create a BETTER RPI RANK than Michigan’s 4 wins over lower ranked teams (Again, IMHO, RPI is DEEPLY FLAWED in this respect).  But they lost all those games which the Committee will certainly make a mental note of…Which takes me back to Michigan Hoops Non-Con scheduling and the need to fill the roster with some beatable mid major 50-100 RPI teams vs. teams over 150 in the RPI.  And also playing another 1 or 2 South Carolina or UCLA’s (RPI top 50) or somebody to have that “quality loss” thing going for you. 

A Valentine Gift tonight with Sparty and OSU squaring off.  Somebody I hate is going to lose…and that’s almost a toss-up regarding Kenpom.  If sparty can’t defend home court….it could be over…

Annnnnd… Sparty then has to go to Purdue and Maryland and knock off Wiscy at home.  Not to mention the always pesky Illini and Cornsuckers are in the mix.  Not a lot of cakewalks there….

That’s not a good look.  Sparty is riding the bubble.