NET Rankings are Flawed

[PSYclone22]

The problem I have with criteria's 2 and 5 is that you are rewarding teams for running it up against sh!t competition. When you look at Iowa's non con it's pretty clear Fran had this figured out because outside of the Cyhawk and ACC challenge Iowa played a pathetic schedule. Notice both of these games were out of his control... I don't care which way you cut it, at some point a win is a win and the efficiency jazz goes out the window. Not all schedules are created equally and that needs to carry some weight.

That's more of a concern for us than anything right now:

It's not clear what percent of NET Rankings is made up of each of the 5 factors. But factors 2-5 do not take into account opponent. We're simply not credited with playing difficult competition twice weekly whereas other teams are credited for 15 point wins against meager opponents. Without knowing how the 5 factors in NET Ranking are distributed I would assume that the factors are (perhaps artificially) capping the number of teams that can make the tournament from one conference, even if they are a deserving team.

 

[cyclones500]

NET Efficiency is just points per possession minus points allowed per possession. That portion of things doesn't appear to take into account opponent.

When it comes to overall NET formula, I have a good grasp, but I'm still hazy about the purpose of the efficiency portion. Is it intended to be a more (ahem) "efficient" way of assessing competitive consistency ... or, as basically a better way to gauge point differential (i.e., how to assess value of a 20-point win vs. close win, for example)? Or is it an additional element of "how good you are"?

 

[BillBrasky4Cy]

When it comes to overall NET formula, I have a good grasp, but I'm still hazy about the purpose of the efficiency portion. Is it intended to be a more (ahem) "efficient" way of assessing competitive consistency ... or, as basically a better way to gauge point differential (i.e., how to assess value of a 20-point win vs. close win, for example)? Or is it an additional element of "how good you are"?

IMO they are trying to mask the fact that win differential matters but everyone is clearly figuring that out since a lot of these efficiency numbers and score differential go hand in hand.

 

[PSYclone22]

The efficiency removes pace from things but is going to be well correlated with the scoring differential. On a scale of descriptive (how well has a team done) to prescriptive (how good would a team do in a neutral match up) it probably goes

Win % --> Adjusted Win % --> Team Value Index --> Scoring Margin --> Net Efficiency

We're merely 76th in Net Efficiency, 75th in Scoring Margin, no idea on Team Value Index and Adjusted Win % (probably our two highest metrics) and around 50-60th in basic Win %.

 

[dahliaclone]

It should be noted that the 'no team with 0 Q1 wins has ever been selected as an at large team' is very likely going to fall. Houston currently has 0 Q1 wins and their only chance to get one is at Memphis at the end of the year. They are 100% in if they don't get a Q1 win.

 

[CychiatricWard]

There is no more weight to offensive efficiency, all points are created equal, it doesn't matter if it's one more point scored or one less point given up.

This can’t be accurate as our defensive efficiency is way better than Iowas so the gap must come from somewhere.

 

[cyclones500]

The efficiency removes pace from things ...

I think that answers a good chunk my question, I had assumed it's a way to avoid putting strict point-difference elements into it.

Random example: 97-70 is a "closer" game than 72-45, even though both are 27 points, based on pace.

That doesn't answer the question of how valuable margin of victory is, but that's a separate rabbit hole.

 

[PSYclone22]

This can’t be accurate as our defensive efficiency is way better than Iowas so the gap must come from somewhere.

Below is efficiency (points per possession)

Team	Off	Def	Net
Iowa	1.145​	0.981​	0.164​
Iowa State	0.965​	0.888​	0.077​

The gap comes from us being MUCH worse on offense compared to them than they are on defense compared to us.

 

[cyclones500]

It should be noted that the 'no team with 0 Q1 wins has ever been selected as an at large team' is very likely going to fall. Houston currently has 0 Q1 wins and their only chance to get one is at Memphis at the end of the year. They are 100% in if they don't get a Q1 win.

Yeah, UH is a lock. Mainly just playing for seed level.

Probably impossible to play its way out. If Houston goes 0-6 to finish regular season, Quads would be:

0-4 5-3 8-2 8-0

(Assumption also 1-and-done in AAC tournament, but don't know opponent, but wouldn't be enough at that point to do much damage).

Granted, if it goes that far south, the seeding drop would be steep.

 

[CychiatricWard]

Below is efficiency (points per possession)

Team	Off	Def	Net
Iowa	1.145​	0.981​	0.164​
Iowa State	0.965​	0.888​	0.077​

The gap comes from us being MUCH worse on offense compared to them than they are on defense compared to us.

I can see that now but it still doesn’t explain away who you are playing. Iowa can beat all the trash teams they want at the bottom of the big ten while scoring 90+ but when push comes to shove they lose when they face a good team. How does being overall more efficient equal a better team? Answer, it doesn’t. We have similar records and beat the breaks off of them yet we are 20 spots lower in NET. Something seems flawed here.

 

[cytor]

All of these stats are enough to make a person hurl. The only stats that matter are Win-Loss and strength of schedule.

 

[MartyFine]

Below is efficiency (points per possession)

Team	Off	Def	Net
Iowa	1.145​	0.981​	0.164​
Iowa State	0.965​	0.888​	0.077​

The gap comes from us being MUCH worse on offense compared to them than they are on defense compared to us.

KenPom's adjusted defensive efficiency rankings ISU is 8th in the country right now. Iowa is 123rd. (Offense Iowa is 5th, ISU 161st). It certainly seems like offense is favored. Of course ISU really struggles with the ball.

 

[MartyFine]

I can see that now but it still doesn’t explain away who you are playing. Iowa can beat all the trash teams they want at the bottom of the big ten while scoring 90+ but when push comes to shove they lose when they face a good team. How does being overall more efficient equal a better team? Answer, it doesn’t. We have similar records and beat the breaks off of them yet we are 20 spots lower in NET. Something seems flawed here.

Yes, KenPom's schedule strength rating has ISU at 23 and Iowa at 82 nationally.

 

[PSYclone22]

KenPom's adjusted defensive efficiency rankings ISU is 8th in the country right now. Iowa is 123rd. (Offense Iowa is 5th, ISU 161st). It certainly seems like offense is favored. Of course ISU really struggles with the ball.

KenPom makes opponent adjustments and is a predictive metric.

I don't think NET does. It's simply raw points per possession.

 

[BryceC]

KenPom's adjusted defensive efficiency rankings ISU is 8th in the country right now. Iowa is 123rd. (Offense Iowa is 5th, ISU 161st). It certainly seems like offense is favored. Of course ISU really struggles with the ball.

No, the margin is favored. Our margin is much smaller than Iowa's.

 

[JM4CY]

KenPom makes opponent adjustments and is a predictive metric.

I don't think NET does. It's simply raw points per possession.

AKA Psy likes it raw

 

[Cyballzz]

There has also never been a team that's more than 4 games under .500 in conference, undefeated in the non-conference, has every team in their conference in the top 70, has no bad losses, and an abundance of tier 1 wins.

There just isn't a direct historical equivalency to our resume this year so being the first team to do something in regards to selection is very possible. Whether it be first team with 8+ quad 1 wins to not make the tournament or first team more than 4 games under .500 in conference to make the tournament.

The closest to it in our history was probably the 91-92 team. Went into the conference tournament at 16-11, 5-9 in the conference. Went 1-1 in the conference tournament.

Went 13-2 in the non con and ended up going into the tournament at 17-12 and was a 10 seed.

Different era of tournament seeding obviously but it would effectively be us getting in with a 7-11 conference record which I would bet solidly gets us in.

I also think the bubble is so bad that 6-12 could be enough but I wouldn't want to take that chance without at least 1 win in KC

 

[NATEizKING]

One thing to start watching. As great as our 8 Q1 wins are, we could be down to half that a week from now.

Creighton is losing at DePaul right now and they're teetering on that Q1/Q2 line as it is.

Xavier is in a slump with a massive tough stretch ahead of them.

Memphis is winning but a loss and they're likely back to Q2

Iowa has two tough games ahead of them...

More likely then end up with more/even:

 

[qwerty]

Looking at some numbers this morning and comparing if the margin of victory were weighted by Quadrant (Q1=1, Q2=.75, Q3=.5, Q4=.25). So a 6 point win in Q1 is equal to 8 point in Q2, 12 point in Q3 and 24 point in Q4.

Iowa State "average" game is ISU 67 opp 62. Average margin of victory is 11.0% (not the 8.1% from average scores). Weighted margin of victory is 2.5%.

Iowa "average" game is Iowa 84 opp 72. Average margin of victory is 19.0% (not the 16.7% from average scores). Weighted margin of victory is 3.8%.

What is really helping Iowa is other than Iowa State game, they have not been blown out by anyone, where ISU has 5 double digit losses. Iowa has 8 weighted margin of victory over 10% but Iowa State has 9. Iowa has 3 losses weighted at more than 10%, Iowa State has 7.

On average, a close loss is not that much worse than a close win, but in reality, it SHOULD all be about the W-L performance. Also, the higher scoring games dilute % against higher numbers, so a 86-95 loss (-9.5%) in metrics is better than a 53-62 loss (-14.5%) although both are 9 point losses.
Due that over an entire season and you start to move off the true metrics of a team's performance.

 

[AuH2O]

IMO they are trying to mask the fact that win differential matters but everyone is clearly figuring that out since a lot of these efficiency numbers and score differential go hand in hand.

Does anyone not understand that net efficiency on a ppp is essentially the same as a scoring margin differential? Your net efficiency and scoring margin are going to correlate almost perfectly.