Top 65 RPI P6 teams

[HRE Otto IV]

If I eyeballed this correctly the only Top 65 RPI Power 6 conference team to NOT make the NCAAT was SU. While TCU was the only P6 to make the NCAAT outside the Top 65.

Our season seems worse than it was because of the highly flawed NET which was used to evaluate everyone all season long. It changed the narrative.

 

[ar inj]

The NET is flawed in the way it allowed teams to game the system, but SU getting blown out on multiple occasions, while probably maybe overstated by the NET, was a black eye on the teams record and indicative of its inability to perform on a consistent basis.

 

[djorange1989]

The NET is flawed in the way it allowed teams to game the system, but SU getting blown out on multiple occasions, while probably maybe overstated by the NET, was a black eye on the teams record and indicative of its inability to perform on a consistent basis.

This

 

[BBQuse]

If I eyeballed this correctly the only Top 65 RPI Power 6 conference team to NOT make the NCAAT was SU. While TCU was the only P6 to make the NCAAT outside the Top 65.

Our season seems worse than it was because of the highly flawed NET which was used to evaluate everyone all season long. It changed the narrative.

Yes, 100%. The RPI is flawed too, but if this was 10 years ago, and we were excluded, this would have been one of the biggest snubs ever. Using some "objective" metrics does make sense, but if there are massive aberrations/outliers, it warrants some common sense stepping in. Winning a lot of close games matters -- it means your team knows how to win. Not scheduling a bunch of cupcakes, so not blowing people out matters. Improving over the course of the season matters.

 

[HRE Otto IV]

The NET is flawed in the way it allowed teams to game the system, but SU getting blown out on multiple occasions, while probably maybe overstated by the NET, was a black eye on the teams record and indicative of its inability to perform on a consistent basis.

No bubble team is consistent. If the NET never existed and the RPI was what the media used all year long to evaluate teams, does SU not make the NCAAT this year? A&M, Northwestern, UVA, Michigan State, Texas, TCU were all behind us.

Both the RPI and NET suck. We should be looking to create something more worthy, especially when it creates the entire narrative for who is worthy or not of an at large bid throughout the season.

 

[Cheriehoop]

The NET is flawed in the way it allowed teams to game the system, but SU getting blown out on multiple occasions, while probably maybe overstated by the NET, was a black eye on the teams record and indicative of its inability to perform on a consistent basis.

This year though very few teams were really consistent. So many teams with some good wins and conversely some bad losses. Heck only UConn, the regular season winners won their conference tournament unlike the Big 10, ACC, Pac 12, SEC, Big 12. There were so many teams that could win and then lose games - many very streaky teams who can look like world beaters one game and absolute crud the next.

 

[jncuse]

You can't use the RPI anymore to evaluate teams, because teams are no longer scheduling in accordance with it. If the RPI Existed, teams wouldn't play as many Q4 teams at home in OOC (and this isn't just the B12, but they are at the forefront) they would revert back to the old way of gaming the RPI - beating mismatched Q3 teams at home.

And trust me the RPI could be gamed just as easily as the NET ... probably more easily... the teams that did best in the RPI were those that could schedule Q3 or teams in the low Q4's that were easy wins at home instead of bottom half Q4 teams. Is beating mismatched Q3 teams at home, better than more mismatched Q4 teams. If I had to choose sure, but it was still a silly way for teams to get ahead of others, when they are all easy wins for a tourney level team.

Note one system punishes Q4 games, and heavily rewards playing Q3 games at home.
The other system rewards Q4 games (since they are easier to control your narrative with blowouts), and doesn't really give any benefit to Q3 games.
A system shouldn't heavily reward winning a bunch of either games.

Hedge them against each other with NET being calculated as (1/2 Current NET + 1/2 RPI) and perhaps teams will be encouraged to play more Q1 and Q2 games in OOC play. But at least the impact of Q4 blowouts would be minimized.

 

[Consigliere]

No bubble team is consistent. If the NET never existed and the RPI was what the media used all year long to evaluate teams, does SU not make the NCAAT this year? A&M, Northwestern, UVA, Michigan State, Texas, TCU were all behind us.

Both the RPI and NET suck. We should be looking to create something more worthy, especially when it creates the entire narrative for who is worthy or not of an at large bid throughout the season.

Here's the difference. The NET describes itself as a predictive based metric. The RPI describes itself as a results-based metric. If you want to award tournament berths based oin what a team accomplished during the season you use one. If you want to select teams based on what they might do in the NCAAs you use the other.

 

[STEVEHOLT]

Here's the difference. The NET describes itself as a predictive based metric. The RPI describes itself as a results-based metric. If you want to award tournament berths based oin what a team accomplished during the season you use one. If you want to select teams based on what they might do in the NCAAs you use the other.

Rewarding a team based on what they may do in the future at the expense of a team that has already accomplished more is beyond stupid. don’t even play the games. Just have the tournament decided by AI (and I don’t mean Allen Iverson)

 

[baggerbob]

Here's the difference. The NET describes itself as a predictive based metric. The RPI describes itself as a results-based metric. If you want to award tournament berths based oin what a team accomplished during the season you use one. If you want to select teams based on what they might do in the NCAAs you use the other.

Using any system based upon what a team might do is stupid.
A player might get sick, cone up with a minor injury, or be out for some other reason.

 

[Millhouse]

If I eyeballed this correctly the only Top 65 RPI Power 6 conference team to NOT make the NCAAT was SU. While TCU was the only P6 to make the NCAAT outside the Top 65.

Our season seems worse than it was because of the highly flawed NET which was used to evaluate everyone all season long. It changed the narrative.

I might be having a brain fart but how can that be with auto bids going to many teams way outside the top 65, I assume

 

[HRE Otto IV]

I might be having a brain fart but how can that be with auto bids going to many teams way outside the top 65, I assume

Power 6 as in ACC, B12, B1G, Big East, P12, SEC. There were other Top 65 teams who didn't make it but were were mid majors.

 

[HRE Otto IV]

Here's the difference. The NET describes itself as a predictive based metric. The RPI describes itself as a results-based metric. If you want to award tournament berths based oin what a team accomplished during the season you use one. If you want to select teams based on what they might do in the NCAAs you use the other.

Well isn't that the point? You picked teams that deserve to be there and not based on what if?


Edit

To expand on this

The NET is used to evaluate teams based on their rankings and the whole quad system throughout the year by media, pundits, and fans. They look at what the team did (results), not what it can do. Which kind of screws up the whole point of having a "predictor" as a metric. What a team actually does against Q1 teams doesn't matter as much as what the predictor says you would do. It matters in a predictor that you play a Q1 team more than the actual result. The predictor can feel you would win by 8 when you actually lost by 2 and "rank" you ahead of that team.

 

[jncuse]

Both the NET and RPI are result based rankings... only difference is one uses margin to measure results.

Calling it "predictive" is because analytical guys have shown across all sports that margin based systems are consistently better at predicting future results. With more and more Q4 games being played across the P6, they might be getting less good at predicting. Not sure. KP (which correlates heavily to the NET) still drives betting lines - if they are no longer predictive the sharps will be doing something. Not sure if they have.

 

[Consigliere]

Well isn't that the point? You picked teams that deserve to be there and not based on what if?


Edit

To expand on this

The NET is used to evaluate teams based on their rankings and the whole quad system throughout the year by media, pundits, and fans. They look at what the team did (results), not what it can do. Which kind of screws up the whole point of having a "predictor" as a metric. What a team actually does against Q1 teams doesn't matter as much as what the predictor says you would do. It matters in a predictor that you play a Q1 team more than the actual result. The predictor can feel you would win by 8 when you actually lost by 2 and "rank" you ahead of that team.

I use this site as my bible:

2024 Men's College Basketball Home Page | WarrenNolan.com

2024 Men's College Basketball Home: NET, RPI, ELO, Scores, Schedules, Rankings, Predictions, Bracket Projection, Simulations, Nitty Gritty, Strength of Schedule, Conference Ranking, Statistics, Streaks

www.warrennolan.com

This is where the predictive-based vs. results-based differentiation is laid out. NET is predictive in that it relies on advanced analytics including offensive and defensive efficiency ratings as keys to its data model - all assumed since the exact algorithms are secret. Based on those data elements it predicts how the team will perform vis-a-vis every other D1 team in future matchups. Yes, results are a part of it as the efficiency metrics are based on games that have been played. It goes a step further however in distilling factors other than what happens on an average possession. It doesn;t matter if a team is up 20 or down 20, it doesn't matter it the game result has already been decided or if we are entering the sixth overtime. It simply considers each possession in a vacuum and evalutes based on that possession and projects how the team will perform based on those data points. The Quad system just stratifies those rankings to provide a snapshot of a team's resume, but a Q1 vs, a Q4 win has very little impact on the actual number. Yes, the committee will evaluate those things, but this is why you seem teams that beat up on low level programs en up so highly regarded in the NET.

RPI is much less complex. It primarily looks at your record, your opponents' record and youe opponents' opponents; record. Second iteration factored in coefficients for home and away. But essentially if you beat a good team by 25 or by 2 there is no difference.