Sagarin over RPI by selection committee

[lawpoke87]

I've been making this argument for weeks. Still not sure why no one else here referenced Sagarin in terms of selection criteria. At least for this year the Sagarin was a far superior indicator.

This post was edited on 3/15 6:08 PM by lawpoke87

 

[URedskin54]

It will be whichever metric fits who the committee already wants to put in. This year it's SOS so Sagarin wins out.

 

[Tulsafanzz]

This year Sagarin favored thr big conferences, so they used it. There were many years it did not favor them & they used RPI.

My take : whatever favors big name schools ( Indiana, UCLA, Texas) is this year's favorite computer ranking choice. Pure garbage. Only Lawpoke & Phoggy think this is fair.

 

[I.I.]

SOS and THE EYE TEST were the biggest factors. Doesn't matter now whether you win or lose just who you play and if you look big and athletic on TV. I'm surprised they didn't get Memphis in with the EYE TEST...they are really good there.

 

[drboobay]

Like I said there is no consistent reasoning year to year. Not a word about body of work and counting early games the same as recent ones.. By the way, TU fails on early and recent so it's not just about Tulsa.

I also don't think the committee is necessarily out to get the little guy. It is more likely unconscious bias. There is a large literature in both economics and psychology that demonstrates purely statistical decision making nearly always beats human, clinical decision making in predicting important outcomes. It is due to peoples inability to overcome biases.
Posted from Rivals Mobile

 

[chito_and_leon]

It's expected TV rating that's the real factor.
Posted from Rivals Mobile

 

[TU_BLA]

I don't buy the TV ratings argument. It's one of the most watched sporting events and fans whose teams are sitting at home are still filling out brackets and watching the first and 2nd weekends. Personally the biggest screw job of the last 30 yrs was theft raking gauntlet Wichita State was forced to endure last year? Does anyone truly think that Kentucky was an 8 seed and Louisville was a 4 last year? No one. Not ESPN, not CBS, no one. I watched a few TX games this year...they're not good. Tons of talent, but they can't play as a team. UCLA doesn't have a ton of talent not do they play well other than a few guys who can shoot. Indiana looked pretty good the games I saw them play so I can't argue much with them.
Posted from Rivals Mobile

 

[cmullinsTU]

If they used Sagarin, I'd have no problem with it. The problem is they are not at all transparent with their process they claim RPI is the only metrics tool they use, which can't be true.

Other than putting UCLA in by name, is they said "well they played Arizona tough for awhile and they're a #1 seed so they can lose well to good teams".
Posted from Rivals Mobile

 

[lawpoke87]

Show me where the NCAA claims that the RPI is the only metric they use in the selection process. In fact, NCAA.org states they specifically use Sagarin, Kenpom, RPI and others in team evaluations. This board may have thought that RPI was the holy grail for whatever reason but the NCAA has always been clear that Sagarin along with other systems are considered.

I predicted in January that the AAC was looking at 2 bids with 3 max depending in Cincy. Sagarin ratings just weren't there. Thought the teams would be SMU and TU with Cincy or Temple fighting for the third.

I'm not sure what's fair. Haven't run the numbers this year yet but the committee has been very consistent in the number of non P5 conference schools in the dance. Guess we'll see if teams like UCLA, Texas and Indiana can compete with higher seeded teams in a few days. If I'm Butler, Texas isn't the 11seed I wanted to see.

 

[URedskin54]

Nobody says it's the holy grail, but the NCAA has said it's the main thing they use. The problem with not sticking to one thing or another is that they can put whoever they want in and then pick any system they want to justify it. Florida managed to play sub-.500 basketball(1-10 top 50) and still end up a tourney team according to sagarin(#42). Clearly schedule strength and point differential is a little overvalued in some cases.

This post was edited on 3/15 9:14 PM by URedskin54

 

[Tulsafanzz]

OU got a great seed & draw. OU a 3 seed...very favorable seed. Albany their 14 seed. Albany is lowered rated (both in RPI & Sagarin), than Tulsa's #7 seed opponent in the NIT, William Mary ! OU needs to give a big shout out to Joe Castiglione, their AD, who just happens to be on the selection committee ! OSU is also an overrated 9 seed. They limped into the tourney. Texas shouldn't be in. Thanks again Joe C. !

Terrible favoritism this year. AAC has no representation on the committee & it shows.

 

[drboobay]

They don't sequester committee members for the entire season. So in the end they can consider whatever they want despite what they might officially discuss in the room.
Posted from Rivals Mobile

 

[lawpoke87]

Please show me where the NCAA says the RPI is the primary tool they us for their selection criteria?

Texas opens as a fav over the 6th seed.

AAC was the 9th rated league. 2 to 3 spots are historically accurate for that slotted league. Stated this all along.

 

[mcane]

Ok state is over seeded and every OSU fan hates this team
and Ford. A donor season ticket holder told me today he hoped they didn't make the tourney.
Another easy money pick of Oregon.
Again this year the number is top 50 games is the
decider.
Not an argument that TU belonged- we had 3
Chances to punch a ticket SMU and Cinncy at
home and perhaps at UConn. I'm ok with the NIT
although I'm blah about the Bill & Hillary team.

 

[drboobay]

Right we had our chances and didn't take them. But seeing UCLA over CSU seems off base. Does not pass the smell test.
Posted from Rivals Mobile

 

[mcane]

Ok state is over seeded and every OSU fan hates this team
and Ford. A donor season ticket holder told me today he hoped they didn't make the tourney.
Another easy money pick of Oregon.
Again this year the number is top 50 games is the
decider.
Not an argument that TU belonged- we had 3
Chances to punch a ticket SMU and Cinncy at
home and perhaps at UConn. I'm ok with the NIT
although I'm blah about the Bill & Hillary team.

 

[ctt8410]

Originally posted by lawpoke87:
I've been making this argument for weeks. Still not sure why no one else here referenced Sagarin in terms of selection criteria. At least for this year the Sagarin was a far superior indicator.

This post was edited on 3/15 6:08 PM by lawpoke87

Because recent history showed RPI was the best predictor for who made it in. SMU was 33 in Sagarin last year and was left out. NC State was 70 in Sagarin and made it in.

They moved the goalposts this year.

 

[URedskin54]

Excuse me, I shouldn't have said "said." Perhaps "implies it by only using RPI data on the data sheets it provides to its members and reveals to the public." Of course other things are used, but what does it imply when the sheet they publish looks like this?

DATA SHEET

 

[lawpoke87]

NCAA.org says they use similar breakdowns from Sagarin, Kenpom, etc. Like I've said for months, some years the selections follow RPI more than Sag while other years it's a Sag bias. The boards fixation with one index puzzled me all season based on historical trends and the statements from the NCAA.

 

[cmullinsTU]

Joe Castiglione said the RPI is the only official metrics given to the selection committee in a recent interview. My post is based on that comment.

Regardless there should be more transparency and honesty about how they make selections. Eye test and recent wins in some years and not others isn't acceptable. I don't think Tulsa earned their way in but its ridiculous that UCLA and Indiana made it in over Temple.
Posted from Rivals Mobile

 

[youngblood223]

Originally posted by cmullinsTU:
Joe Castiglione said the RPI is the only official metrics given to the selection committee in a recent interview. My post is based on that comment.

Regardless there should be more transparency and honesty about how they make selections. Eye test and recent wins in some years and not others isn't acceptable. I don't think Tulsa earned their way in but its ridiculous that UCLA and Indiana made it in over Temple.
Posted from Rivals Mobile

This wasn't metric based. Probably a coincidence that Sagarin justified some of this BS. UCLA isn't even in a play-in game. There's no justification for it.

 

[lawpoke87]

Per Mich State AD Mark Hollis the committee members are given breakdown sheets from the KPI, Sagarin, KenPom, LRMC and BPI of all RPI top 100 teams. They are given these sheets every Monday for the entire season and when they convene for final selections. This coincides with what's posted on the NCAA website. Not sure what Joe C was talking about. Anybody else shocked the BPI breakdown sheet is now being used by the committee?

 

[ctt8410]

Originally posted by lawpoke87:
Per Mich State AD Mark Hollis the committee members are given breakdown sheets from the KPI, Sagarin, KenPom, LRMC and BPI of all RPI top 100 teams. They are given these sheets every Monday for the entire season and when they convene for final selections. This coincides with what's posted on the NCAA website. Not sure what Joe C was talking about. Anybody else shocked the BPI breakdown sheet is now being used by the committee?

Do you have a link to that comment because it contradicts everything I've read and heard about the selection process.

 

[ctt8410]

Originally posted by lawpoke87:
Please show me where the NCAA says the RPI is the primary tool they us for their selection criteria?

Greg Shaheen was on the pregame show for UConn-SMU and stated that the committee prefers RPI because they would rather not broach the philosophical discussion raised by a margin-of-victory system. More importantly, the NCAA publishes RPI daily on their website, posts RPI selection sheets for every team, and puts on a mock selection exercise that media members have written about extensively in opposition of how much they emphasize RPI:

"R.P.I.'s fingerprints were all over the process. When a computer monitor displayed the teams that we were considering for the bubble, the R.P.I. ranking was listed suggestively alongside them. The color-coded "nitty gritty" worksheets that the committee has developed, and which often frame the discussion about the bubble teams, use the R.P.I. rankings to sort out the good wins and the bad losses."

http://www.nytimes.com/2012/03/11/sports/ncaabasketball/rpi-may-belong-outside-the-ncaa-tournament-bubble.html?_r=0

"Every "nitty gritty" page the NCAA uses, every fact and figure and list of top 50 wins and strength of schedule and noncon SOS and you name it is broken down based on RPI[/I]. You can't sit in the selection room and not[/I] be affected by RPI. It underpins every consideration the committee makes, whether the committee always knows it or not."

http://espn.go.com/blog/collegebasketballnation/post/_/id/48853/ncaas-rpi-line-as-tired-as-ever


"The NCAA allows (but from what I interpreted, does not heartily endorse) any Selection Committee member to use Massey or any type of ratings system (including - WHAT - theCoaches' Poll[/I]? It's true, unfortunately). Those systems are not brought up on the big screen, unless by request, which never happened at our mock.
I'm guessing its seldom a Pomeroy team page will get clicked to the projector screen this year, too, especially since it's now subscription-based, and that $20 annual fee might be a touch too much.
The RPI is the peanut butter that keeps the primary data smudged together for the Selection Committee. It still permeates the process. And the NCAA still wants to deny that. The NCAA likes to say bringing up a team's specific RPI ranking doesn't often come up when debating two teams' inclusion or seeding. While that's true, from the outset, the organization, presentation and general data on a team is dressed up in an RPI shirt with an RPI hat and a cute pair of RPI gloves."

http://www.cbssports.com/mcc/blogs/entry/26283066/34887185

 

[Gold*]

Thank you, ctt.

 

[HHOOTter]

Mike Sorensen: NCAA committee overrates strength of schedule By Mike Sorensen, Deseret News
Published: Sunday, March 15 2015 7:55 p.m. MD

SALT LAKE CITY - One quibble I have every year with the NCAA tournament selections - and I know many others share the same view - is that the tournament committee places too much emphasis on strength of schedule.

Strength of schedule is overrated because many schools can't control who they play and teams shouldn't necessarily be rewarded just because they lost to a bunch of good teams.

NCAA selection committee chairman Scott Barnes, Utah State's athletic director, talked about strength of schedule in justifying why UCLA and Texas made this year's field of 68.

"(UCLA) was one of the tough decisions the committee had to make, but they did have a good strength of schedule,'' Barnes said on CBS Sunday afternoon.

Of Texas, he said, "Strength of schedule was the determining factor in that.''
BYU should have been a higher seed than both of those schools, which both avoided the play-in game, while the Cougars must play Tuesday against Mississippi (7:10 p.m. MT, TruTV). But the Cougars didn't have the luxury of a built-in schedule like the Bruins and Longhorns had in their respective conferences.

It stands to reason that schools from big conferences have better strength of schedules because of their conference slates, which they're forced to play. Schools in mid-major conferences don't often have the opportunity to play big-name teams, who don't have much to gain by scheduling lesser teams that may beat them.

Instead of the Bruins and Longhorns, who get in every year, why not put in a team from a mid-major conference like Old Dominion or Richmond or Murray State, which went unbeaten in the Ohio Valley Conference, but was upset in its conference tourney?
Sure, you can argue that UCLA and Texas might beat Old Dominion and Richmond, but I think it's more fun seeing more mid-major teams in the tourney than some of the same old teams
+++++++
Bids by conference
Big Ten: 7
Big 12: 7
ACC: 6
Big East: 6
SEC: 5
Pac-12: 4
A-10: 3
Mt West: 3
American: 2
MVC: 2
WCC: 2

12 "@ Large" Bids went to nonP5 Conferences
w/ "Media Darling" Big Least landing 5

Regardless of the RPI/Sagrin/Strength Index
I don't think it's right for a team that didn't even finish .500
in the regular season in their conference to ger a NCAA bid.
(OSU & Texas)

On play in games,
Instead of 4 play in games,
why not go to 16 play games
for the #13 seed to #16 seeded teams?
That way u would eliminate the other
dozen or so "Bubble" teams,
& all those boarderline bubble teams can claim a bid!

Personally, might as well extend the NCAA tourney one week
& have the top 64 teams host a 1st rd NCAA game @ their home court
128 teams would make the field
& conference Tourney champs would B given a home game.
Imagine, schools like Hampton, Lafayette, or Robert Morris hosting a first rd game
vs. the like's of Indiana, Texas, OSU, 'ol Miss, or UCLA?



This post was edited on 3/16 8:10 AM by HHOOTter

 

[TU_BLA]

Don't forget the NCAA is under a lot of heat from the P5 schools. Ignoring 2 super power teams (Texas and UCLA) to put in a Temple and Colorado St. would not endure the NCAA to the P5 schools in a time when the NCAA is fighting for its very existence. Remember, its the schools that employ the NCAA and none of them are obligated to stay within that structure. The P5 schools and the TV contracts for NCAA events because of those P5 schools is what keeps the NCAA operational. The P5 schools could up and leave and form their own little Wall St Collegiate Association and leave the G5 and mid-majors to the NCAA and the NCAA would be gone in 5 years.

 

[youngblood223]

Originally posted by TU_BLA:
Don't forget the NCAA is under a lot of heat from the P5 schools. Ignoring 2 super power teams (Texas and UCLA) to put in a Temple and Colorado St. would not endure the NCAA to the P5 schools in a time when the NCAA is fighting for its very existence. Remember, its the schools that employ the NCAA and none of them are obligated to stay within that structure. The P5 schools and the TV contracts for NCAA events because of those P5 schools is what keeps the NCAA operational. The P5 schools could up and leave and form their own little Wall St Collegiate Association and leave the G5 and mid-majors to the NCAA and the NCAA would be gone in 5 years.

I don't think that has as much to deal with it as TV audience for those games.

 

[lawpoke87]

I'm not going to link the NCAA.org site again as I linked it in the other thread. It clearly states the list of rating indexes considered which are considered.

Here is the text from the article which I was referring. Thought it gave some honest insight you don't get from the NCAA talking heads:



EAST LANSING - Things might have worked out differently for Kevin Pauga if not for the superior oblique palsy in his left eye, which required two surgeries at age 12, led to scoliosis in his back and effectively ended any chance of a high school sports career.

Then again, maybe not. Long before that, he was a 5-year-old reading the TV Guide, fascinated by the patterns of scheduling, color coding it for his own enjoyment.

"These are things," he said, "a 5-year-old shouldn't do."

But this was not your typical 5-year-old. And this is not your typical 32-year-old. This is one of the emerging names in college basketball analytics, one that may soon be invoked by hoops geeks everywhere with as much frequency and affection as Ken Pomeroy and Jeff Sagarin.

The NCAA men's basketball selection committee convened Wednesday at the Conrad Hotel in Indianapolis to start putting together the NCAA tournament bracket that will be unveiled Sunday, and the rankings of Pomeroy, Sagarin and Pauga are among the data they'll consider.

A formula Pauga started tinkering with as a Michigan State senior in 2004 was finally published and shared with the public as the KPI in 2013, and now it is cited regularly by Sports Illustrated, ESPN.com and the like.

It has grown into college football rankings and other analytics at KPISports.net, but it started as a kid who wanted to improve on the RPI - the Ratings Percentage Index, developed by the NCAA in 1981 and still used as the committee's primary vehicle for sorting information about teams.

"I saw some major flaws in the RPI," Pauga said. "So what I tried to do in my 22-year-old, punk-ish way was try to come up with a better solution."

The RPI assesses a team taking into account its winning percentage (25% of the ranking), the winning percentage of its opponents (50%) and the winning percentage of its opponents' opponents (25%). It has been tweaked over the years and, for example, now counts a road win as 1.4 wins and a home win as 0.6 of a win.

The KPI ranks the games, each team's profile consisting of a growing log from best performance to worst. Each night, he enters that day's results into a 60-megabyte Excel file that he calls the "matrix" - it's around 5,600 games this season and counting.

The KPI uses opponent winning percentage, opponent strength of schedule, scoring margin, pace and location in its rankings. And the adjustments for home/road/neutral are set by the data, not pre-determined - in other words, if a college basketball season ever saw road teams win more than home teams, home wins would count for more than road wins.

"It's become one of the favorites in the room," MSU athletic director Mark Hollis said of the 10-member selection committee and the KPI, which Hollis brought to the committee in 2013 after encouraging Pauga to make it public.

For the 2013-14 season, the committee used the KPI as a discussion point. This season, it has become part of a breakdown sheet each Monday delivered to committee members, comparing the RPI top 100 with KPI, Sagarin, Pomeroy, BPI and LRMC rankings.

"It's genius," David Worlock, NCAA director of media coordination and statistics, said of Pauga's formula. "And as a stats junkie myself, it's admirable - it's not like he has a staff."


It's obvious to anyone whose followed the selection process and seedings closely over the past ten years that the committee doesn't follow any of these indexes as it's primary source year in year out. Rather it's a blend of those systems and personal preference (have no other explanation for some of the selections). There are studies out there (one from Harvard) which illustrates what a poor predictor the RPI is as far as predicting games compared to KenPom and others. Probably a discussion for another thread.

Remember....the NCAA basketball tourney is the only sporting event from which the NCAA receives the revenue. I'm not naïve enough to think TV ratings from matchups and teams in the field don't play a role. The NCAA has a vested interest unlike any other college game or event.

 

[cmullinsTU]

Originally posted by lawpoke87:


It's obvious to anyone whose followed the selection process and seedings closely over the past ten years that the committee doesn't follow any of these indexes as it's primary source year in year out. Rather it's a blend of those systems and personal preference (have no other explanation for some of the selections). There are studies out there (one from Harvard) which illustrates what a poor predictor the RPI is as far as predicting games compared to KenPom and others. Probably a discussion for another thread.

Remember....the NCAA basketball tourney is the only sporting event from which the NCAA receives the revenue. I'm not naïve enough to think TV ratings from matchups and teams in the field don't play a role. The NCAA has a vested interest unlike any other college game or event.

I bolded the portion of your comment that irks me the most about the selection process. It seems to an outsider that they select the index that justifies their selection the best. I also agree that TV ratings are a large part of the decision making, although that will never be admitted by the NCAA.

 

[lawpoke87]

RPI, Sagarin and KenPom have been around awhile and have proven themselves over time to be a useful tool in evaluating teams. I found it interesting that ESPN comes up with their own system (BPI) and it's immediately included in the handouts given to the committee members.

 

[JesseTU]

We look at how a team closes a year.

We look at overall performance.

We look at RPI.

We use an eye test.

We think strength of schedule is really important.

We think whatever justifies getting what we want is important.

There is no criteria. Its a crap shoot. What pisses me off is that the result is more for the haves. When was the last time a non P5 got an at large bid in place of a seemingly more deserving "big time" team? The criteria is simple:

"Whenever you can make an argument for it, always pick a P5 team over a midmajor. Always."

The current trend would be to schedule as strong of a schedule as possible, then win 1/3 of the hard games, clean up on lower teams, and hope you have a big name to carry the day. Right Indiana? UCLA?

Also, don't be in the American conference, or it seems you have a good shot of being screwed every year.

(This is not am argument for Tulsa to be in this year)

 

[lawpoke87]

Here's a link to KenPom's ratings. Note that the top 44 teams all made the tourney this year...impressive to say the least. He's been hit and miss in the past but for this year he was very accurate.

http://kenpom.com/index.php

 

[ctt8410]

Originally posted by lawpoke87:

Here's a link to KenPom's ratings. Note that the top 44 teams all made the tourney this year...impressive to say the least. He's been hit and miss in the past but for this year he was very accurate.

http://kenpom.com/index.php

It (or at least the components that influence it) was clearly a large factor this year, which means that the blowout loss to UConn hurt more than we realized. We peaked at 55 after beating Houston then dropped to 79 after that loss.

Here are KenPom top 40 (regular season) that were left out last year:

32 SMU
35 Louisiana Tech
36 Utah
38 St. John's