Plaino something happened in the RPI this week that shows RPI bias against OU

[bullmarket]

Explain this Plaino Baylor at home beat Ok State RPI #36 in the first two games of their series and Baylor at home lost game three.. In the new RPI ratings Baylor went to #10 from #9.
Baylor therefore lost one place in the RPI ratings

Compare this to LSU playing at home was playing #31 Missouri
LSU like Baylor won the first two games over Missouri at home but like Baylor LSU lost game three at home.
But LSU gained one place in the RPI ratings passing OU..

BOTH LSU AND BAYLOR WHO WERE PLAYING SIMILARLY RANKED OPPONENTS AT HOME WITH THE EXACT SAME RESULTS LOSING GAME 3 AT HOME.

BUT in the RPI Baylor lost one place FALLING from #9 to #10 in the RPI while LSU GAINED one spot passing OU in the RPI.

LSU GAINED one place in the RPI while Baylor LOST one place in the RPI when each of these home teams had identical series vs. opponents identically ranked.

WHY DID LSU GAIN A SPOT TO PASS OU IN THE RPI RANKINGS WHILE BAYLOR LOST A SPOT IN THE RPI RANKING WHEN IDENTICAL THREE GAME SERIES AND IDENTICAL RESULTS OCCURRED FOR BOTH BAYLOR AND LSU VS IDENTICAL OPPONENTS

This shows RPI bias against OU letting LSU pass OU gaining a spot in the RPI when Baylor fell one spot in the RPI

Equal occurrences for two teams having identical results should have identical results in the RPI ratings. But Baylor and LSU had opposite results in the RPI . because the RPI committee wanted to harm OU

This is proof of RPI's absolute anti OU bias because in a computer generated program when an identical occurrence occurred the computer would have generated an equally identical result in the RPI rankings UNLESS A HUMAN BEING INTERVENED TO CHANGE THE RESULTS WHICH OBVIOUSLY HAPPENED FOR LSU TO PASS OU IN THE RPI RANKINGS.

 

[anon_5z68cgtvj4ggn]

 

[Plainosooner]

Bull, I don't think RPI can be biased per se. I think it could have a bad program basis. But I thought the explanation in another thread was solid, in understanding what's going on. It seems a little flawed. But we haven't played as many quality teams and beat them as most of the others in the top 16.

 

[Josh_M]

So much stupid packed into one thread - not aimed at you Bull or Plaino.

So many deletions.

Two bans.

Nothing has changed here guys. Be smart.

 

[JConXtsy]

Uh, Baylor dropped a spot because Tennessee played #1 Florida which boosted their Opponents' Winning % a bunch, so Tennessee jumped them regardless of losing 2 to Florida. Once you get up to 50 games in a season, losing a couple of games doesn't really affect your Winning % much, but Opponents' Winning % can still fluctuate a lot.

 

[bullmarket]

Uh, Baylor dropped a spot because Tennessee played #1 Florida which boosted their Opponents' Winning % a bunch, so Tennessee jumped them regardless of losing 2 to Florida. Once you get up to 50 games in a season, losing a couple of games doesn't really affect your Winning % much, but Opponents' Winning % can still fluctuate a lot.

JConXtsy, Plaino & Syberite (Our 3 RPI experts) please explain this to me.

When OU beat Baylor 2 out of 3 including a huge 6-0 OU win Baylor was ranked #10 & OU #12.
Here is what the RPI did after the OU & Baylor series..
Baylor despite losing 2 out of 3 to OU moved UP to #9 and OU remained at #12

Please explain the math or the logic of RPI's actions especially how could.Baylor move up after losing 2 out of 3 to OU and why didn't OU move past Baylor?

Thanks..

 

[iasooner1]

bull

Like JCon explained earlier, the RPI is like a computer; an unthinking mathematical equation.

 

[bullmarket]

bull

Like JCon explained earlier, the RPI is like a computer; an unthinking mathematical equation.

Yet the highly respected College Softball Coaches poll has OU ranked #6 and Baylor #18

Its not like the human poll and the computer poll have very similar rankings. OU is ahead by a mile in the human poll & OU is behind Baylor in the RPI computer

iasooner my great friend the "RPI computer" is broken.

 

[sybarite43]

A computer operates under the formula that it is given. It is not capable of determining the sanity of that formula.

 

[JConXtsy]

When OU beat Baylor 2 out of 3 including a huge 6-0 OU win Baylor was ranked #10 & OU #12.
Here is what the RPI did after the OU & Baylor series..
Baylor despite losing 2 out of 3 to OU moved UP to #9 and OU remained at #12
.

Actually, both Baylor and OU moved up a spot after that series. From what I'm looking at, OU was #13 and Baylor was #10. Baylor moved up to #9 and Oklahoma moved up to #12.

April 16, 2017 RPI Rankings
https://extra.ncaa.org/solutions/rpi/Stats Library/SB Nitty Gritty thru 4_16_2017.pdf

April 23, 2017 RPI Rankings
https://extra.ncaa.org/solutions/rpi/Stats Library/SB Nitty Gritty thru 4_23_2017.pdf

Again, you're reading into the Wins/Losses too much at this point in the year. When you've got nearly 50 games in the books, winning and losing doesn't shift your overall Winning% too much right now. However, SOS' are in great flux when you consider you're playing such a huge range of teams (1-150+).

As Sybarite pointed out, home/away wins/losses are weighted as well. Home wins are only 0.7 pts. Away wins are 1.3 pts. So the Baylor/OU series in Norman would have 2 wins going to OU (2 * 0.7 = 1.4 pts), and 1 win going to Baylor (1.3 pts). Basically, Baylor going 1/3 in Norman was almost a win/loss wash for both teams.

 

[JConXtsy]

I want to say this to you, Bull, about the RPI. You're not the first person to find umbrage with it. It's well known to be a poor indicator of future performance, and it's controversial to say the least.

Rankings that take into account the "severity" of wins/losses have been proven to be far better indicators of future performance. However, those rankings were eventually deemed inappropriate for sportmanship sake, since they tended to encourage score padding.

 

[bullmarket]

I want to say this to you, Bull, about the RPI. You're not the first person to find umbrage with it. It's well known to be a poor indicator of future performance, and it's controversial to say the least.

Rankings that take into account the "severity" of wins/losses have been proven to be far better indicators of future performance. However, those rankings were eventually deemed inappropriate for sportmanship sake, since they tended to encourage score padding.

So JConXtsy when the Regionals are decided who will host how much role do the two respected polls (The Coaches Poll and the ESPN Poll) play in the Regional decisions?

 

[JConXtsy]

So JConXtsy when the Regionals are decided how much role do the two respected polls (The Coaches Poll and the ESPN Poll) play in the Regional decisions?

Honestly, I don't follow softball enough to know. I tried searching for information confirming Plaino's comments that the committee mainly uses the RPI for selection, but I couldn't find any verification of that comment at all.

This guy predicts Oklahoma hosting a Norman Regional as of 5/1/2017.
http://www.collegesportsmadness.com/softball/bracketology

Looking at his resume', he apparently writes quite a bit about D1 softball. Writing doesn't make you know, but it shows some level of commitment and following.

 

[Oklabama]

Yet the highly respected College Softball Coaches poll has OU ranked #6 and Baylor #18

Its not like the human poll and the computer poll have very similar rankings. OU is ahead by a mile in the human poll & OU is behind Baylor in the RPI computer

iasooner my great friend the "RPI computer" is broken.

Bull, I don't get it completely either. I try not to worry about it though because it gives me a big headache without even having a gin. So I've just decided to let it all play out and watch the girls win as many as they can knowing that if they don't lose, we have another National Championship. How great would that be!!! Boomer!

 

[Plainosooner]

Honestly, I don't follow softball enough to know. I tried searching for information confirming Plaino's comments that the committee mainly uses the RPI for selection, but I couldn't find any verification of that comment at all.

This guy predicts Oklahoma hosting a Norman Regional as of 5/1/2017.
http://www.collegesportsmadness.com/softball/bracketology

Looking at his resume', he apparently writes quite a bit about D1 softball. Writing doesn't make you know, but it shows some level of commitment and following.

I hope he's right, but he's got us ahead of both UTenn and Washington, who've both played a lot tougher schedules and both have wins over us, and Oregon, who is ahead of us in every poll. The Ducks are 5, we're 6 in the coaches' poll. The Ducks are 5, we're 7 in the ESPN poll. And the Ducks are 7, we're 13 in the coaches poll.

The RPI isn't the only consideration. But since the voted polls are just opinions, they don't rate as high as the computer as those giving their opinions in the voted polls. I keep pointing back to two years ago, but we were ranked in the voted polls top eight, but were the 11 seed, mostly because of RPI.

I believe that the Baylor issues will be resolved by the committee, if we win the regular season and post season tournament. But they have more wins against top ranked teams, and if we falter in Bedlam, or if Baylor wins the Big XII post season tournament, then I'll bet they are seeded ahead of us. But since they're likely to be third in the conference, if we win both, I'd be shocked if they're seeded ahead of us.

The RPI doesn't really consider head to head stuff for tie breakers. It's more about what the computer thinks of your schedule. But if you getting credit for the record of the teams you play, and not the teams they play, then the best schedule would be to play the best teams in the minor conferences, who'll post a bunch of wins.

This format really encourages that. And I think playing Tulsa is part of OU's strategy. Heck, Tulsa is in the top 25. They alone don't bring you up, but enough like that is good for your RPI.

The credit for head to head, and for conference winners comes from the committee in the bracketing.

 

[roygbell]

I'm sure not an expert in these kinds of issues. My guess is that the rpi is just a simple number derives from a complex computer generated algorithm. The algorithm is built to take into consideration the whole of the season, and not based on individual head to head matches. It tries to compare all the match-ups of the whole season. It takes in all of a teams records and compares it to all of the records of the rest of the teams. The result may not seem to be logical as it is based on mathematical calculations that may have unintended built in bias.

In the early part of this season, OU wasn't playing great while playing a difficult schedule. We lost quite a few games by very small margins. We have lost 8 games. 6 of the games [Auburn, Washington, Tennessee, ND, Arizona and Baylor] were one point losses. Two others losses were to Cal Poly by scores of 3 - 1 and 3 - 1. So, when you look at those losses and with OU playing a fairly weak conference schedule doesn't make us look all that good.

My guess is that the algorithm compares the losses to Auburn, BYU, Washington, ND, Tennessee and Arizona heavily against us. And, we really don't have any really great wins to counter those losses, that is unless you want to count UCLA and the two Baylor wins.

 

[Plainosooner]

We beat BYU soundly. Some of these programs include run differential. Some don't. On the RPI, it doesn't seem to matter.

 

[bullmarket]

We beat BYU soundly. Some of these programs include run differential. Some don't. On the RPI, it doesn't seem to matter.

Sixteen teams host a regional That we will do.
The Sixteen regional winners play in the Eight Super Regionals
Win our regional at home and we are in a Super regional.

With a little luck we could host the Super in Norman,.

The fraudulent RPI will have no effect.

 

[Plainosooner]

It will have some. But I think we will be ahead of at least a couple who are ahead of us if we win the conference, and the post season tournament. That rain out last weekend could keep Tech out. Without KU playing as many as everbody else, both ISU and KU could both finish ahead of them.

 

[sybarite43]

It will have some. But I think we will be ahead of at least a couple who are ahead of us if we win the conference, and the post season tournament. That rain out last weekend could keep Tech out. Without KU playing as many as everbody else, both ISU and KU could both finish ahead of them.

Kansas 0
Iowa State 1

Texas 1
Baylor 2

Arizona 0
UCLA 8

Tennessee 1
Texas A&M 5

Utah 0
Oregon State 5

Kentucky 2
Mississippi State 6

LSU 4
South Carolina 5

Minnesota 11
Penn State 1 (6)