https://www.cbssports.com/college-baske ... g-the-rpi/

The catchy thread title is my one contribution. I'll let the Number Nerds do the heavy lifting.

If you search "RPI" on Ramblermania you get 57 pages of hits.

But no replies to RPI's demise, guys? Surprised.

I just don't know what to think. The optimist in me says this is a great thing since we are typically an efficient team and more times than not advance stats rank us higher than the RPI. But the critic in me sees some flags.

They're telling us some things that they're using but not the weights. They are still using the Quadrant System. If that is a heavy weight like it seemed to be this year, then I don't think much will change. It doesn't fix the scheduling issue where Power teams refuse to play us and similar mid-majors. Your top non-power conference SOS teams were Temple (13) Wichita St (29) UCONN (32) and Illinois St (54). https://www.cbssports.com/collegebasketball/rankings/sos My favorite line in the article is "In addition, the NET will also tweak some data to 'ensure fairness.'" To me, that sounds like "we can adjust things to get the results we want"

Finally, I'm a little worried that they don't release what it'd look like in previous years or even what the model is. Everyone knows that you can't change selections of the past. So either they are hiding it because they don't want to admit that they totally screwed up, or they think it wouldn't have changed anything and want people to think for at least a year that this new thing will solve problems, similar to how the Quadrants were supposed to make things fair last year.

So I'm just waiting to see how it plays out. Hopefully it helps, but I'm always skeptical when it comes to the selection process. They always seem to find ways under each system to prevent really good mid-major schools from getting in while making sure that average power schools get a bid.

There's a discussion on this over at MVCfans.com: http://mvcfans.com/viewtopic.php?f=2&t=5052

One thing that concerns me is metrics like the Strength of Record, and using winning margin. If you read through the series on Cinderella teams, you'll see that scheduling (certainly) and winning margins (probably) are going to help major conference teams. Those plus metrics for major conference teams are going to show up in the NET, which will continue to use the Quadrant system developed last year as a seeding/at large metric.

So basically, it seems like not much has changed, except adding a performance metric (margin of victory) and instituting other metrics like Strength of Record to substitute for strength of schedule.

To me, based on what I saw from last year under the Quadrant system, the main metric should be won-loss record (not just victories, but the won-loss ratio of all games ) against the top 100. You could expand or shrink that top 100 number to fit the top 2 Quadrants and it would still probably work. However, I think there will be some years where the imaginary cut-off line between good and not-so-good teams will be 67 and some years it will be 110. In other words, the quality and parity among good teams is sometimes wide and shallow, sometimes narrow and deep.

The committee is throwing out recent performance and making it more open to counting early performances (supposedly in the name of giving more credit to mid-majors who do well in non-conference), but I don't believe that's ever been a problem. Seems to me it's giving cover for teams like Oklahoma last year who played well early but stunk it up late in conference. I think record in the last 15 games (including late conference season and conference tournament) should be a secondary metric, but considered. Two of the Final Four teams last year had 14 and 13-game winning streaks going into San Antonio, which meant that neither team had lost since late January/early February.

So won-loss percentage against top 100 or top 2 quad teams should be the main measure, after the NET (the replacement ranking from the RPI). In other words, if there are two teams with a NET of 35-39, you'd take the team that's 4-1 (.800 winning percentage) against top 100/top 2 quad teams over the team that's 5-7 (.417 winning percentage). In the recent past, the excuse has always been "they've got five wins against top 100 teams!", ignoring that they've also got seven losses.
Also, if a team is 12-1 in their last 13 games, you'd bump them a little higher than a similar team that's 8-5 down the stretch.

So I don't see how anything has changed much. They're not going to take a team like last year's St. Mary's over Oklahoma, Syracuse, Alabama, or the other teams that got in with an under-.500 conference record. That would be the first thing I'd ban (the winning percentage among top 100/Top 2 quad teams should take care of that). There should be a downside to big money conferences hoarding good programs, propping up mediocre programs, and excluding/squelching/condemning potentially good programs-- that's what they're doing now. If you say that games against good teams are the only ones that really matter, but you're still taking teams that lose most of their games against the good teams they know best, then you're not being honest, realistic or fair.