The Gophers RPI is:

Win on the road against P5 competition and most all of the analytics will love you..All except RPI.
 


Why does anyone care about RPI, it will not and should not be used.
 


Kind of sucks that this is the season the NCAA gave up on RPI as it would benefit the Gophers and the rest of the B1G a great deal this season if they still used the RPI as no other conference comes even close to having as favorable of an overall RPI as the B1G has. And it's strange to me that they went from basing their decisions so heavily on the RPI to dropping it completely, but outside of how it would benefit the Gophers, I say, GOOD RIDDANCE, and it's about frigging time!!! RPI has to be one of the worst analytic rating systems ever devised.

The NCAA never did base a team's individual RPI that heavily. They did use teams records vs Top 25, Top 50 Top 100 etc, but only among other criteria such as Overall record, Road/Neutral Record, SOS, SOS nonconf etc...

They had changed the record vs RPI stuff last year with the advent of Quads 1, 2, 3 & 4.
 
Last edited:


Of course the RPI has flaws, and of course many of the different analytic systems are better for predicting future performance, however, if you believe that seeds should be “earned”, then at some point a win has to count clearly more than a loss. A 1 point win should not count as a 51% (in simplified terms), nor should a 1 point loss count as a 49%, even in 2018.

Now, if the NET is used simply as a tool (as RPI supposedly was) to bucket teams for evaluating strength of wins and losses, then that sounds good. We’ll see this year.
 

RPI still matters because it's what the quads are based upon.

Also, I would argue if you are top 40 RPI and from a major conference, you will make the tournament.
Still and easy non-biased equation to determine if you are tournament level or not.
 

RPI still matters because it's what the quads are based upon.

Also, I would argue if you are top 40 RPI and from a major conference, you will make the tournament.
Still and easy non-biased equation to determine if you are tournament level or not.

Not true. NET ranking is what will be used for quads as well.

https://www.sbnation.com/college-basketball/2018/8/22/17768922/ncaa-tournament-selection-process-rpi-net-metric-committee

Current NET Rankings

https://www.ncaa.com/rankings/basketball-men/d1/ncaa-mens-basketball-net-rankings
 

RPI still matters because it's what the quads are based upon.

Also, I would argue if you are top 40 RPI and from a major conference, you will make the tournament.
Still and easy non-biased equation to determine if you are tournament level or not.

Better do some research.
 





What the hell...
When did this happen?

I take a half season off of college basketball, and suddenly the whole world is changed.

Our RPI is 10.
Our NET is 52???

How the heck are we 52???

Can anyone provide how the NET ranking is calculated?
 

What the hell...
When did this happen?

I take a half season off of college basketball, and suddenly the whole world is changed.

Our RPI is 10.
Our NET is 52???

How the heck are we 52???

Can anyone provide how the NET ranking is calculated?

Has NCAA updated NET rankings from last night? The bookmark I have is still showing rankings only through games played Jan. 2.

Have a hard time believing NET ranking would remain the same (that's what it was through Wednesday's games) after a road win at the #15 team.
 



Has NCAA updated NET rankings from last night? The bookmark I have is still showing rankings only through games played Jan. 2.

Have a hard time believing NET ranking would remain the same (that's what it was through Wednesday's games) after a road win at the #15 team.


Not official, but this has us at 52 through Jan 2nd.

http://warrennolan.com/basketball/2019/net


Down 7 spots from Dec 30th.

I assume they just copy the NCAA.
https://www.ncaa.com/rankings/basketball-men/d1/ncaa-mens-basketball-net-rankings
 


Has NCAA updated NET rankings from last night? The bookmark I have is still showing rankings only through games played Jan. 2.

Have a hard time believing NET ranking would remain the same (that's what it was through Wednesday's games) after a road win at the #15 team.

If NET is some version or using KenPom as an input, the 52 rating has to have something to do with last year's data being carried over to this year? The last ten games last year were a disaster. When and/or if those numbers are diluted and gone, it should help the Gophers.

Pitino had an interesting comment on pre-game presser that most of the analytics don't care about winning. Strictly efficiency data. I suppose one could be really efficient but playing more efficient teams and lose, but seems unlikely.
 

If NET is some version or using KenPom as an input, the 52 rating has to have something to do with last year's data being carried over to this year? The last ten games last year were a disaster. When and/or if those numbers are diluted and gone, it should help the Gophers.

Pitino had an interesting comment on pre-game presser that most of the analytics don't care about winning. Strictly efficiency data. I suppose one could be really efficient but playing more efficient teams and lose, but seems unlikely.

His point is that a 1 pt win and a 1 pt loss are treated virtually the same because the analytics only care about the aggregated per possession outcomes, rather than W/L. While reasonable for predicting future performance, or quality of an opponent, I continue holding out hope that the actual outcomes againt great/good/decent opponents (i.e records v. tiers and other reasonable considerations) is what will carry the day in the end.
 

His point is that a 1 pt win and a 1 pt loss are treated virtually the same because the analytics only care about the aggregated per possession outcomes, rather than W/L. While reasonable for predicting future performance, or quality of an opponent, I continue holding out hope that the actual outcomes againt great/good/decent opponents (i.e records v. tiers and other reasonable considerations) is what will carry the day in the end.

+1. Future prediction is helpful for fans. However, rankings used to determine the NCAA tournament should be based on results.
 

Gophers are up to 34 in NET after today’s update
 


It appears the formula consists of 5 different areas.

1. Team Value Index (probably similar to current RPI. Based on wins losses and strength of opponent with road, home, neutral site figured in)
2. Net Efficiency - points scored per possession - points allowed per possession
3. Winning percentage
4. Adjusted winning percentage - Reflection of RPI formula that was used in the past.
5. Scoring margin

Items 1, 3, and 4 all seem to be items that the current RPI system covered, or could have been modified (as it was in the past) to better match the goal of what they wanted.
Items #2 and #5 are new and it's hard to tell the impact / weighting of each.
 

In looking at the rankings, some interesting things. NET Ranking seems most closely aligned with Net Efficiency. One outlier however is Houston who is ranked #4 in NET rankings but lower in Net Efficiency.
However they have won 8 of their 12 games by 10+ points, so this seems to be the biggest reason they are a team out-performing their NET efficiency.

Houston:
NET =4
RPI = 32
Efficiency = 12
Margin 10+ games: 8 of 12 wins

Gophers
NET 34 / was 52
RPI: 10
Efficiency: 49
Margin 10+: 5 of 12 wins

Interesting thing about the margin of victory is that it's probably redundant with Net Team Efficiency, so the weight of it might be pretty heavy.

NET Ranking

1 1 Duke ACC 11-1 0-0 4-1 7-0 0-0
2 2 Virginia ACC 12-0 2-0 3-0 7-0 0-0
3 3 Michigan Big Ten 14-0 2-0 2-0 10-0 0-0
4 4 Houston AAC 14-0 2-0 0-0 12-0 0-0
5 6 Gonzaga WCC 13-2 1-1 3-1 9-0 0-0
6 8 Texas Tech Big 12 12-1 1-0 3-1 8-0 0-0
7 5 Tennessee SEC 11-1 1-0 2-1 7-0 1-0
8 9 Nevada MWC 14-0 3-0 4-0 7-0 0-0
9 7 Michigan St. Big Ten 12-2 2-1 2-1 8-0 0-0
10 11 Kentucky SEC 10-2 1-0 1-2 8-0 0-0
34 52 - Minnesota

Of the top 10 NET teams, all are top 13 for Net Efficiency, except Houston.

Net Efficiency:
Rk Team
1 Duke
2 Virginia
3 Gonzaga
4 Michigan St.
5 Michigan
6 Nevada
7 North Carolina
8 Virginia Tech
9 Kansas
10 Texas Tech
11 Tennessee
12 Auburn
13 Kentucky
14 Nebraska
15 Purdue
16 Iowa St.
17 Florida St.
18 Florida
19 Wisconsin
20 North Carolina St.
21 Mississippi St.
22 Ohio St.
23 Cincinnati
24 Oklahoma
25 TCU
26 Villanova
27 Texas
28 Indiana
29 Maryland
30 Buffalo
31 Creighton
32 Houston

49 Minnesota



RPI Team
1 Kansas
2 Oklahoma
3 Nevada
4 Duke
5 Tennessee
6 Michigan State
7 Buffalo
8 Michigan
9 Florida State
10 Minnesota
11 TCU
12 Houston

Margin of Victory
Wins by 10 points or more
Houston 8 of 12 wins by 10 points or more
Minnesota has 5 of 12 wins by 10 or more.
 
Last edited:

Utah jumps up to 96 in NET with their big win over Arizona ST and Santa Clara moves up to 188 with their win over San Diego pushing them into a tier 3 victory as opposed to a 4
 

Utah jumps up to 96 in NET with their big win over Arizona ST and Santa Clara moves up to 188 with their win over San Diego pushing them into a tier 3 victory as opposed to a 4

Santa Clara is on fire. 8-6 now with 6 straight wins.
 

It appears the formula consists of 5 different areas.

1. Team Value Index (probably similar to current RPI. Based on wins losses and strength of opponent with road, home, neutral site figured in)
2. Net Efficiency - points scored per possession - points allowed per possession
3. Winning percentage
4. Adjusted winning percentage - Reflection of RPI formula that was used in the past.
5. Scoring margin

Items 1, 3, and 4 all seem to be items that the current RPI system covered, or could have been modified (as it was in the past) to better match the goal of what they wanted.
Items #2 and #5 are new and it's hard to tell the impact / weighting of each.

Thanks for posting this. I'd assumed (faulty) it was basically just #2. I have a tough time with #5, as anyone who's gambled on basketball games knows, the final margin can look a lot different that the margin with 2 minutes to go. Unless the math works out as a sort of bonus for winning by over 10, and a neglibible difference for margins of 1-9. I could get behind that as a valid factor.
 

I don’t understand why the selection committee would consider rankings that are so dependent on efficiency. Yeah, efficiency helps determine which teams really are the best and worthy of top 25, but for March Madness I think teams’ resumes as far as wins and losses should be the deciding factor (obviously with SOS taken into account). The margin of a team’s win shouldn’t really be considered nor should advance efficiency stats
 

In looking at the rankings, some interesting things. NET Ranking seems most closely aligned with Net Efficiency. One outlier however is Houston who is ranked #4 in NET rankings but lower in Net Efficiency.
However they have won 8 of their 12 games by 10+ points, so this seems to be the biggest reason they are a team out-performing their NET efficiency.

Margin of victory is capped at 10 for NET rankings.
 

Santa Clara is on fire. 8-6 now with 6 straight wins.

And Josh Martin is playing well for them. They have an excellent coach in Herb Sendek. Maybe that's not as bad a win as we thought. With a few quality teams in the WCC, they might even be able to stay under 200.
 

I don’t understand why the selection committee would consider rankings that are so dependent on efficiency. Yeah, efficiency helps determine which teams really are the best and worthy of top 25, but for March Madness I think teams’ resumes as far as wins and losses should be the deciding factor (obviously with SOS taken into account). The margin of a team’s win shouldn’t really be considered nor should advance efficiency stats

Why ? Check out Kenpom at the end of he season and it will show if your efficient on offense and defense ppp you will be wildly successful with wins, every year this holds true.
 

If NET is some version or using KenPom as an input, the 52 rating has to have something to do with last year's data being carried over to this year? The last ten games last year were a disaster. When and/or if those numbers are diluted and gone, it should help the Gophers.

Pitino had an interesting comment on pre-game presser that most of the analytics don't care about winning. Strictly efficiency data. I suppose one could be really efficient but playing more efficient teams and lose, but seems unlikely.

Well if you do great in Kenpom those teams win conference titles and win 25 or more games all the time. Last years games have dropped off Kenpom so they are a reflection of how you play and against whom. Look at the top 10, who thinks who should not be there ?
 




Top Bottom