Keeping an eye on the national rankings

[DownGoesKansas]

We jumped Illinois in NET this morning to 36th, but they play Ohio St tonight - so unless they get killed, they'll probably jump 2-3 spots.

Especially because it is on the road

[CyCarp]

We jumped Illinois in NET this morning to 36th, but they play Ohio St tonight - so unless they get killed, they'll probably jump 2-3 spots.

[sevelev711]

Back to my original post #425 asking for clarification on the NET: many PN posters have said our net efficiency is much better this year because of the improvement in our offense. If so, how come there is such a large discrepancy between NET and RPI and 10 Big 10 teams are ahead of us in this metric?

Because RPI cares more about wins and losses, which will benefit a team that's 25-5. But if a team that wins by an average of 20 compared to an average of 2 against the same schedule, you'd certainly have a decent argument that one team is better than the other. With RPI that's not the case, it just compares win-loss. NET compares efficiencies. And our offense is actually really good this year, 18th in the country (according to Kenpom), but our defense is not that great, 93rd in the country. Compare that with the #1 team according to Kenpom, Kansas, who is 7th in offense and 2nd in defense. That's where the discrepancy comes from. RPI, which only cares about win-loss of you and your opponents, is higher on us than efficiency metrics, which lean more on how you play, not necessarily the results.

[Panther AC]

This is partly true but winning by 40 does count more than a 10 point win. One of the inputs it includes is margin of victory that is capped at 10 points. Another input is net efficiency (adjusted offense - adjusted defense) which almost by definition can't be capped.

Back to my original post #425 asking for clarification on the NET: many PN posters have said our net efficiency is much better this year because of the improvement in our offense. If so, how come there is such a large discrepancy between NET and RPI and 10 Big 10 teams are ahead of us in this metric?

[UNIAlum2010]

For the MOV input it "hurts" us is that way that Iowa's road loss to Purdue was capped at 10 instead of the full amount but you also have to remember that we got big boosts in the NET for blowing out Drake/Missouri State on the road. It would have been even bigger if that part wasn't capped.

[CyCarp]

I'd be curious to see the impact of limiting MOV to 10 points. It obviously hurt us, since we have no double digit losses. It seems like every Big Ten team has lost at least a game or two by 20+, although they win games by 20+ too. It just amazes me how inconsistent those teams are.

[UNIAlum2010]

Including margin of victory is what is really helping P5 teams. Purdue is a good example. They have lost a lot of games on the road to "good" teams. Losing by a few points on the road a top 50 team means you are essentially a top 50 team. Add in a few blowout wins and you get a huge boost. Winning is still more important than Kenpom/Sagarin etc but less important than RPI was.

[sivert]

The NET is probably* better than RPI in accurately assessing the strength of teams based on the schedules they have played.

The problem (IMHO) is the schedules played. P5s want to play lots of home games against cupcakes ($$$). The RPI didn't care about Margin of Victory. It cared about: W-L record and SOS. So when they beat cupcakes(bad for SOS) at home(bad for SOS), they only got credit for a "W." Didn't matter that they won 80-30. NET adds efficiency and limited MOV.

So the RPI should have encouraged more P5 teams to play away or neutral games against quality teams. If they had done that, the limitations of the RPI would have been negated. But they didn't want to schedule like that. They want to play at home. And the NET lets them do that without it hurting them.

I think that schools should push for a limit on the number of home games each school plays.

* "Probably" because it does have the AI component, which nobody really knows what that is trying to do. It has been said that the AI is simply seeking to predict wins in the NCAA tournament. In that case the best thing to do is for Mid-Majors to win in March at a higher clip than other bubble teams that got in.

[Trapped in California]

This is partly true but winning by 40 does count more than a 10 point win. One of the inputs it includes is margin of victory that is capped at 10 points. Another input is net efficiency (adjusted offense - adjusted defense) which almost by definition can't be capped.

I should have been more clear about the margin of victory measurement being capped in for that measurement only. The NET thing is a plate of spaghetti - hard to separate one thing from another.

[DownGoesKansas]

Plus the P5 adjustment within the black box piece.

[UNIAlum2010]

Net includes measurements on how efficient the offense and defense are against the opponents. Margin of victory counts, to a point. Winning by 40 supposedly counts the same as winning by 10 in that metric.

This is partly true but winning by 40 does count more than a 10 point win. One of the inputs it includes is margin of victory that is capped at 10 points. Another input is net efficiency (adjusted offense - adjusted defense) which almost by definition can't be capped.

[Trapped in California]

Net includes measurements on how efficient the offense and defense are against the opponents. Margin of victory counts, to a point. Winning by 40 supposedly counts the same as winning by 10 in that metric.

[Panther AC]

Was looking at NET vs. good ol’ RPI today.

RPI
UNI 19
B1G teams ahead of UNI 1 (next closest is at 32)

NET
UNI 37
B1G teams ahead of UNI: 10

Weird, I wonder why they switched to NET?!?

Would some knowledgeable PN poster please explain, in a sentence or two, the difference in metrics factors between RPI and NET? In other words, what does NET include that RPI doesn't, to have such a discrepancy?

[DownGoesKansas]

Does the committee even look at RPI anymore?

I think just for SOS. Which helps a little but not as much as actually using RPI

['70 Grad]

On Sunday night's "Big Ten & Beyond" show, the feature on bubble teams had Mike DeCoursey (sp) singing our praises as a team that would get his consideration if we didn't get the automatic qualifier. Had a nice highlight package of the Drake game. Steven Bardo's pick was E. Tenn. St. No others were mentioned.