History of KPI

Submitted by Gabranto on

I saw on the front page Brian mentioned that the KPI was the only bracket on Bracketmatrix that had us at the 5 seed...

I did a little research and saw this on their site and found: "Information on KPI Sports includes a compilation of data, analytics, algorithms and analysis from Michigan State University Assistant Athletic Director Kevin Pauga."

Just thought the board would like more conspiracy theories.

Here is the link to learn more about their history: http://www.kpisports.net/about/

Indy Pete - Go Blue

March 8th, 2018 at 11:03 AM ^

"As leagues grow through expansion and strength of schedule grows in importance, the ability to compare two teams who many not play becomes more difficult."

Yet, it has MSU over UM - which did play each other TWICE.  To quote Shooter McGavin, 'Who won that fight any way?'  And yes, KP did misspell 'may' by inadvertently putting 'many'.

Please visit the link given by OP - read Kevin Pauga's musings.  This index sounds like a 12 year-old kid who had an idea for science fair that finished in 2nd place only b/c the science teacher plays cards with his dad.  I am extremely unimpressed by his description of his own index.

This must have been added when Mark Hollis was in power - he has the Minus touch (not to be mistaken with the Midas touch - which Warde Manuel seems to have)

bronxblue

March 8th, 2018 at 2:08 PM ^

KPI Sports has come to grow into more than just a 50 MB Excel file with literally millions of cells of data.

You'd think people who claim to understand advanced statistics would know about the existence of freaking relational databases. It's insane to me that people have 50 MB Excel files as their primary date store and analytical toolset. This alone should disqualify KPI from being taken seriously.

1989 UM GRAD

March 8th, 2018 at 9:42 AM ^

In his defense, he does also have MSU as a 4 Seed.

I don't feel like spending any time at his site, but I'm guessing most of the seed projections are based on some sort of analytics or formula...which likely take the personal element out of it.

Then again, I could be wrong.

Mike Damone

March 8th, 2018 at 10:34 AM ^

in conspiracy theories.

I am a believer that a Sparty could be intellectually challenged when it comes to math and algorithms.  Don't know why anyone would be surprised at a poor end product here. 

If he was predicting the quality of the upcoming corn crop in the Midwest, would say his Sparty background would be of value...

TrueBlue2003

March 8th, 2018 at 2:38 PM ^

Kevin Pauga has no statistics education or background, he built this in a freaking excel spreadsheet.

He is a completely unqualified hobbyist that built a random model that, is not only way off the predictive models, it appears to simply attempt to predict how the committee would seed teams back when they factored in top 50 wins and NCSoS, both of which have been de-emphasized or removed as factors.

Totally garbage model and insane that it's part of the team sheets.

 

Hold This L

March 8th, 2018 at 10:47 AM ^

Next week I'm working on a formula to rank every bowler in the world. Hint: It won't make any fucking  sense.

Arb lover

March 8th, 2018 at 11:16 AM ^

I would never have spent the time digging into who KPI was, but it made me chuckle to find out. Thanks for this.

I don't understand why some would neg what appears to be decent and relevant research/information that helps Michigan's cause, unless they were not Michigan fans or didn't like the title or something else like it was posted 3x already (which it hasn't been). Gabranto's obviously been a long time lurker/Michigan fan but only rarely posts, and posts quality. 

ThatTCGuy

March 8th, 2018 at 11:32 AM ^

I don't think this is some big conspiracy or anything by the Staee athletic department, I don't think that athletic department is competent enough for that. I just think that they're terrible at coming up with algorithims.

SpartyJon

March 8th, 2018 at 1:08 PM ^

who made that index. The only one I pay a lot of attention to is Ken Pom. Does any of this matter though? The committee is using RPI topped with RPI sprinkles.

Edit: totally in left field, but TV Ted is officiating the Virgina/Louisville game and just made a bad call and to top it off made it really late.

Indy Pete - Go Blue

March 8th, 2018 at 1:21 PM ^

KPI is on the team sheets being used by the committee.  It is one of a handful of metrics being used to seed teams.  The fact that it appears to have the quality of a blind koala throwing darts is legitimately concerning.  Logic like yours and metrics like KPI are the only way an inferior team like MSU gets seeded ahead of UM.

aMAIZEinBLUEinTX

March 8th, 2018 at 1:52 PM ^

Has anyone glanced at the team sheets?  I've been perusing them throughout the last month; I believe these are the final tangibly printed metrics that the committee uses in scrubbing the seeds on the S-Curve...

The latest iteration is March 6 (sorted by current RPI ranking) -

https://extra.ncaa.org/solutions/rpi/Stats%20Library/March%206,%202018%…

Michigan (RPI: 13 KPI: 20, SOR: 5, AVG: 12.67) (BPI: 14, POM: 9, SAG: 11, AVG: 11.33) ( AVG: 12)

(Michigan has since moved to 13th on BPI, to average 11 in the 2nd equation/ 11.83 overall)

The laughably debunked KPI is not only the 2nd number prominently displayed in the top left corner of the team sheet, but also factors into 2 condensed formulas (of 3 and 6 variables) that the committee will use to ultimately give Staee Detroit, and send us to Wichita or San Diego.

 

Utter crap that a '20' has to be front and center on our team profile