Massey vs. RPI rating and commentary on UCF vs. UConn

[RockyMTblue2]

Antonelli is an ACC homer and an SEC syncophant and it poisons pretty much everything she touches. If I have to watch a game she's doing I hit the mute button quite frequently. If she's on the Louisville game it will be bad, very bad.

 

[stamfordhusky]

Do you guys know Ken Pomeroy or Kenpom, a rating system used in MCBB?

I know him well. Great site. I wish he also did his analysis for the women. I think he does a more accurate job than does Massey, and he also produces dozens of interesting analyses.

 

[buddy]

ESPN has to support the ACC members over UConn. They have invested a lot of money in that conference and they need it to succeed. Thus, the announcers must always try to promote ACC teams over anyone else.

 

[KnightBridgeAZ]

As I mentioned in my previous post, RPI places twice as much consideration on the record of your opponent than it does on your own W-L records. So even though Stanford lost its game, it gained even more by adding a 17-1 team to its list of opponents.

As I also mentioned, hypothetically a team could go 0-30, losing every game by 50+ points and be a top 10 RPI team if it played the right schedule. Losing to teams like Utah is a boon to the RPI.

Repeating doesn't make it completely true. If you went 0-30, by definition you would have a 0 RPI factor.

Oddly enough, the result of the single game you win has the greatest single effect on your raw RPI number. An opponent winning a single game (never mind diluting it to your opponents opponents which is 25% of your RPI score) is very diluted. It does make a difference who you played, but their individual wins / losses don't have a huge effect.

And as others said, Massey is predictive. RPI is not predictive on a game by game basis and is not intended to be. And no one in the business thinks it is.

Playing teams that have a good winning percentage always helps your RPI, but Stanford's change is probably due just as much to what others did around them than adding Utah to the mix.

 

[stamfordhusky]

Repeating doesn't make it completely true. If you went 0-30, by definition you would have a 0 RPI factor.

Oddly enough, the result of the single game you win has the greatest single effect on your raw RPI number. An opponent winning a single game (never mind diluting it to your opponents opponents which is 25% of your RPI score) is very diluted. It does make a difference who you played, but their individual wins / losses don't have a huge effect.

And as others said, Massey is predictive. RPI is not predictive on a game by game basis and is not intended to be. And no one in the business thinks it is.

Playing teams that have a good winning percentage always helps your RPI, but Stanford's change is probably due just as much to what others did around them than adding Utah to the mix.

In this case, I repeated because it IS true. From current teams' records, I could easily construct a scenario where a team could have lost every game by over 50 points and be ranked in the top 10 of RPI. That assumes a team is an independent and gets to make their own schedule. Going 0-30 by definition only produces a 0 RPI for 25% of a team's total RPI. The RPI for the other 75% in theory might be 80% - which would produce an overall RPI of 60%. Etc.

As for Stanford, I don't have time to do the math, but I'd be very surprised if adding a 17-1 record (Utah) to the overall record of its opponents - which counts for 50% of its RPI - wasn't a bigger factor than the random wins and losses its previous opponents had over a one or two day period.

 

[Bigboote]

If she said UConn has an RPI in the 30's, then she was even more clueless than usual. Current UConn RPI is #7.

One big difference between RPI and models like Massey is, as others have pointed out, that the RPI only looks at whether you win or lose - not the margin of victory. In the RPI, a one-point win is the same as an 80-point win. Further, the RPI evaluates the strength of the teams you beat solely by their records, not by who they beat or lost to and by how much. Also a team gets less credit for its own wins than it does for the wins of its opponents. If a team had a W-L percentage of 0.0% (i.e. no wins for the entire season), played opponents with a W-L percentage of 75%, who in turn played opponents with a W-L percentage of 60%, its RPI (ignoring home/road factor) would be 52.5% - well above average.

In theory a team could have a record of 0-30 and lose every game by 50+ points, and end up ranked in the top 5 in the nation in RPI. It's not going to happen - but it could, if a team was able to put together a schedule that only included road games vs the seven or eight power-conference teams with the best W-L percentages.

During Antonelli's delivery, there was a graphic up that had UConn at #7. I think the OP had a typo. They're projected to be a little lower by the end of the season. Antonelli pointed out, factually, that no team with an RPI above 7 had been a #1 seed. I think pretty much anyone would agree that if they win against Louisville and South Carolina, there's no question that they're a #1.

In addition to your points about the RPI, there's also the fact that beating a #100 team counts more than beating a #300 team. And if you look at, say, NC State vs UConn, the equalizer is that NC State has played a bunch of teams in the 70-100 RPI, whereas UConn has played quite a few with RPI greater than 100. I think smart teams are gaming the system with their schedules, scheduling a bunch of teams they think will be in the 50-100 range.

 

[Bigboote]

In this case, I repeated because it IS true. From current teams' records, I could easily construct a scenario where a team could have lost every game by over 50 points and be ranked in the top 10 of RPI. That assumes a team is an independent and gets to make their own schedule. Going 0-30 by definition only produces a 0 RPI for 25% of a team's total RPI. The RPI for the other 75% in theory might be 80% - which would produce an overall RPI of 60%. Etc.

There was one team earlier in the season (Northwestern State maybe?) that was something like 1-10, but had played a bunch of top-ranked teams, and they were, IIRC, in the top 50 in RPI, so I'm with you. It can become ridiculous.

 

[Plebe]

This times 1,000! Amin drives me nuts (I’m not a fan of Beth Mowins either....). During the Oklahoma game he must have said “they are doing this against the number 1 team in the country” about 30 times. I kid you not. Go back and watch it. It was 30 times. Annoying as explitive.

Uh, the UConn-Oklahoma game was on FS1. Adam Amin works for ESPN.

 

[KnightBridgeAZ]

As for Stanford, I don't have time to do the math, but I'd be very surprised if adding a 17-1 record (Utah) to the overall record of its opponents - which counts for 50% of its RPI - wasn't a bigger factor than the random wins and losses its previous opponents had over a one or two day period.

Stanford's RPI went down. They also played and beat a dreadful Colorado team. RPI dropped from .6983 to .6910 Stanford's RPI rank went up, from 4 to 3.

They moved up past Oregon, who played the dreadful Washington schools this past week and whose RPI fell from .7020 to .6894 and whose ranking dropped from 3 to 4. In spite of winning both games.

You are not conceptually wrong - but there are a lot of factors involved. Most importantly, the rankings themselves are based on very small numbers changing. And it can be very volatile - Ohio State went from #129 to #99 to #117 in the last 3 weekly rankings. The actual RPI of course didn't change more than .02

All this from RealTimeRPI.

 

[KnightBridgeAZ]

I think smart teams are gaming the system with their schedules, scheduling a bunch of teams they think will be in the 50-100 range.

Of course they are. Back in the late 90's, one of the WBCA magazines (my wife is a member) had an "anonymous" article about how to game the RPI. I always thought it sounded like Pat Summitt, but then again she wouldn't have had to worry about it, she played genuinely good teams in the day.

Really good (for their conference) teams are desirable, because they will rack up a lot of wins in conference. Really poor P5 teams are good, as well. You get all the benefits from their conference opponent's seasons in factor 3, since most P5 teams have a strong winning percentage OOC.

And don't play teams that you (theoretically) can't beat, unless they are anticipated to be very good. Losses to RPI builders don't necessarily look good, losses to top 10 and 25 teams are no problem.

And there was more, but I've forgotten it.

I'll add you can't get carried away trying to game the system, you do need some real wins on your resume to get a good seed in the NCAA's, and a good RPI without some quality wins is not going to do anything for you.

 

[hussybulldog]

She just implies that UConn’s Coach is not the best as his nation’s best 5 starters have not won all the games.

But didn't she also say, she believes UConn has the best 5 starters in the nation?

 

[maccca3232]

At present UConn - RPI 7, SOS 30, Record 18-1.

In last 12 seasons.

Teams with top 50 SOS with less than 3 losses occurred 34 times.

1 Seed x 32
2 Seed x 2
Other x 0

Notables:
Baylor in 2017-18 - RPI 4, SOS 23, Record 31-1, Seed 2
Duke in 2012-13 - RPI 5, SOS 25, Record 30-2, Seed 2

Teams with top 50 SOS with 3 losses occurred 24 times.

1 Seed x 8
2 Seed x 14
Other x 2

Notables:
Ohio St. in 2006-07 - RPI 11, SOS 43, Record 27-3, Seed 4
Middle Tenn. St. in 2006-07 - RPI 16, SOS 49, Record 28-3, Seed 5

 

[Bama fan]

So...all of you reaffirmed by belief that Antonelli is a knucklehead, RPI is a severely flawed stat and if we take care of business, we have nothing to worry about.

One last point that Antonelli needs to realize:
WCBB is NO DIFFERENT than the rest of the NCAA Sports on how early rounds are conducted on the top seeds campuses. Softball and Baseball take it one step further and have the Superregional (16 to 8) also at the 1-8 seeds campus. Why? Because of attendance!
She’s delusional on her views of what WCBB is vs. the reality of what occurs come the Sweet 16 sites.
There are no “neutral courts” in the SouthEast or out West that either are close enough to their local WCBB elite team or are willing to bid to host.
Baylor has been sent to both Dallas and OKC and attendance was abysmal.
Sweet 16s usually aren’t on Campus except for Rupp Arena and attendance at both years of their hosting was the worst of the 4 sites and one year ND was there!
Can we send her and Adam Amin off to purgatory?

The entrance requirement for purgatory is such that one must suffer in order to expiate their sins. Some sinners are judged to have committed such grievous transgressions that no expiation is possible. They end up a bit further down. One such grievous sin is disrespecting UConn WBB. I fear for the soul of Ms. Antonelli. She must atone at once.

 

[SimpleDawg]

This is why this RPI nonsense used to discredit a team is very infuriating. RPI doesn't tell the entire story. And it's a calculation that's focused largely on a few aspects and biased for a team that happens to come across a string of favorable situations.

Mississippi State is being discredited for the same reason UConn is. That's just degrading given that anyone who watches this game knows that every time those 2 take court, the entire audience feels that a slaughterhouse is coming, regardless of opponent.

While our schedule seems to have very few bad apples left (6 out of 9 opponents in the RPI top 55), UConn still has 8 opponents remaining ranked RPI #95 or lower with 4 RPI group fours. Expect this RPI talk around UConn's schedule to increase.


....

 

[PAHuskeyFan]

I admit, I have not paid a lot of attention to RPI...and I have NO intention of doing more research to figure it out... but I find it totally and completely unreasonable to expect that:
a. A Team loses to an opponent with a worse ranking in every poll,AP, Coaches, Massey and RPI, but yet their RPI improves. AND it's their 2nd loss of the season.
b. A Team loses to an opponent (on the road) NOT ranked in either AP or Coaches Top 25 including "receiving votes", ... and this Team had previously lost another game (at home) to a team with a worse RPI, BUT they retain the #1 RPI ranking, even though they drop in the AP, and Massey and will likely drop in Coaches poll.

Sorry... I find none of this logical, reasonable or proper way to "run an airline," as they used to say... It makes my brain hurt! Back when I was in high school (more than a few years ago, I admit), the State High School athletics program didn't have a State Playoff system for Football. There was an Eastern Region and Western Region Champion...then a "VOTE" to decide the mythical "State Champion." I am so VERY THANKFUL that the NCAA Women's Basketball Championship will be decided on the Court and not by using nonsensical polls such as RPI. If it doesn't make sense to the "average WBB Fan".. it shouldn't be used in making decisions such as Tournament ranking... Just sayin!!!

 

[Sluconn Husky]

But didn't she also say, she believes UConn has the best 5 starters in the nation?

She said she believed UConn's starting five was the nation's best offensive unit.

 

[MSGRET]

This times 1,000! Amin drives me nuts (I’m not a fan of Beth Mowins either....). During the Oklahoma game he must have said “they are doing this against the number 1 team in the country” about 30 times. I kid you not. Go back and watch it. It was 30 times. Annoying as explitive.

I believe that was actually against Notre Dame where he was talking about CW and how she was scoring against them as a freshman.

 

[UcMiami]

Just to be c;lear on what RPI means: the index comprises a team's winning percentage (25%), its opponents' winning percentage (50%), and the winning percentage of those opponents' opponents (25%).

Teams and conferences have learned that you can game this system - the best way is to play the best teams in the weakest conferences - you get a win which ads .25 to your RPI, and because you beat a team that will win 90% of its conference games regardless of how it does in its OOC, you get .50 times at least those 14 wins. And even better you get into your conference with even the bottom have of your conference having 8-12 wins on their resume which you get 50% rpi credit for when you blow them out.

The bottom four teams in the ACC arrived at conference play with a combine record of 39-13 (RPI gold for the rest of the league) and have now gone 2-26!

5 of the 6 bottom teams in the Big 12 when 46-7 OOC and are now 13-25 (Sherri at OK didn't do her job compiling a 4-7 OOC

The most honest conference actually appears to be the SEC in terms of OOC more closely matching conference record which is a change from past years.

(SOS: The opponents' winning percentage and the winning percentage of those opponents' opponents both comprise the strength of schedule (SOS). Thus, the SOS accounts for 75% of the RPI calculation and is 2/3 its opponents' winning percentage and 1/3 its opponents' opponents' winning percentages.)