NET Rankings

[XUFan09]

So X clobbers a so-so Miami of Ohio, and Auburn and Wisconsin (2 of X's 3 losses) both get to the NET top ten after Ohio State's loss, and presto-chango, X leaps from 118 to 62. It seems to me that NET is a bit hyperactive, over-reactive when given relatively few data points to work with. Time will tell if it produces reasonably accurate rankings by the end of the season. At present, both the KenPom and Sagarin rankings look more reasonable to me.

Both Kenpom and Sagarin include preseason weights that drop off as more games are played. If they didn't, they would be hyperactive right now too.

[GumbyDamnit!]

I think what's funny about all this data crunching this early is that our focus seems to be on the 2nd and 3rd degrees of separation. Instead of wondering if a Miami loss to Rutgers is good or bad for specific BE programs based on head to head results, maybe at this point we should just recognize that a Miami loss might simply be good for all at year's end. They're probably in that 7-12 range of ACC teams--clearly in that usual bubble range for this conference. This loss might end up hurting them tremendously should Rutgers finish in the basement of the B10 and they are on the cusp of a bid.

Just win the games in front of you and let the data take care of itself. Clearly the NET is going to wildly fluctuate until more games can be played anyway.

[stever20]

I think what's funny about all this data crunching this early is that our focus seems to be on the 2nd and 3rd degrees of separation. Instead of wondering if a Miami loss to Rutgers is good or bad for specific BE programs based on head to head results, maybe at this point we should just recognize that a Miami loss might simply be good for all at year's end. They're probably in that 7-12 range of ACC teams--clearly in that usual bubble range for this conference. This loss might end up hurting them tremendously should Rutgers finish in the basement of the B10 and they are on the cusp of a bid.

Just win the games in front of you and let the data take care of itself. Clearly the NET is going to wildly fluctuate until more games can be played anyway.

I think this is a great point. But also you have to if you're playing a bad team play well vs those teams.

However looking at Miami- they're 5-2 so far. Last 5 OOC games are pretty putrid- lowest win % chance is a game @ Penn- with 73% chance. So probably will be 10-2. Looking at their ACC schedule it's brutal. 11 of their 18 games vs top 50 teams. So if they can get to 9-9 in conference play, they're going to be safe. What the loss vs Rutgers did was probably take the chance of them getting in at 8-10 off the table....

I said it prior to the game- Miami losing to Rutgers would be good for the Big East. Why? Because it makes Rutgers better- and for St John's- they need it. St John's got Rutgers on the road- so for Rutgers to be a tier 2 game- the threshold is only top 135. For Seton Hall- with their great OOC schedule- their selection is going to be only on record. If they can get to 18-19 wins, they're making the tourney. Miami being good really is meaningless in that....

[gtmoBlue]

I think what's funny about all this data crunching this early is that our focus seems to be on the 2nd and 3rd degrees of separation.

Just win the games in front of you and let the data take care of itself. Clearly the NET is going to wildly fluctuate until more games can be played anyway.

Amen brother.

[adoraz]

Stever, I don't know how you can look at these numbers and not see that they clearly overvalue beating up on cupcakes. I'm stating more of a fact here rather than an opinion. It is not just me saying this, it's also analytics experts like Nate Silver (538 founder).

A team like Houston absolutely should not be ranked #34. You're either being naive or dishonest in defending them. Their best win is @ #130 (NET) BYU. Their other wins are all against awful competition at home. You theorize away games being a major factor, but no that's not it. Houston only has ONE away game and ZERO neutral games. In other words, their schedule sucks big time to date. Their schedule gets a lot better in December, but as of today it has to be one of the worst in the nation.

Why do you think the AAC is doing so much better in NET than RPI? It's because the AAC's SOS is horrible. They hardly participated in the tournaments last week, and they haven't had a conference challenge.

No need to be argumentative here, what I'm saying is actually a good thing for the AAC. They may have a better year due to scheduling a bunch of cupcakes. Don't worry, I'm not the NCAA. I won't change it up. I actually prefer it this way being a SJU fan.

How is it that the NCAA overlooked this? I have no clue.

[stever20]

Stever, I don't know how you can look at these numbers and not see that they clearly overvalue beating up on cupcakes. I'm stating more of a fact here rather than an opinion. It is not just me saying this, it's also analytics experts like Nate Silver (538 founder).

A team like Houston absolutely should not be ranked #34. You're either being naive or dishonest in defending them. Their best win is @ #130 (NET) BYU. Their other wins are all against awful competition at home. You theorize away games being a major factor, but no that's not it. Houston only has ONE away game and ZERO neutral games. In other words, their schedule sucks big time to date. Their schedule gets a lot better in December, but as of today it has to be one of the worst in the nation.

Why do you think the AAC is doing so much better in NET than RPI? It's because the AAC's SOS is horrible. They hardly participated in the tournaments last week, and they haven't had a conference challenge.

No need to be argumentative here, what I'm saying is actually a good thing for the AAC. They may have a better year due to scheduling a bunch of cupcakes. Don't worry, I'm not the NCAA. I won't change it up. I actually prefer it this way being a SJU fan.

How is it that the NCAA overlooked this? I have no clue.

Well, several AAC teams had tournaments the week before which count just as much..... UCF, UConn, Wichita. Pretty much I think everyone but Houston, ECU had a tournament.... So that goes out the door. And tournaments the week before count exactly the same as tournaments last week....

I do think Houston is somehwat overrated right now- but that after tomorrow will change after they see Oregon.

I think one thing that is gone is the thought you can play a great team and lose, and have it benefit you.....
Providence right now has the #139 SOS. A LOT of that is obviously playing Michigan. The 2 teams they lost to are 8-3. The teams that they beat are 12-13. I think their SOS in the NET reflects a lot more the 12-13 part of the schedule instead of the 8-3 part of the schedule. Part of why they are much worse in the NET than they are the RPI. They're looking a lot more at the SOV than they are the SOS.

[adoraz]

Stever, I don't know how you can look at these numbers and not see that they clearly overvalue beating up on cupcakes. I'm stating more of a fact here rather than an opinion. It is not just me saying this, it's also analytics experts like Nate Silver (538 founder).

A team like Houston absolutely should not be ranked #34. You're either being naive or dishonest in defending them. Their best win is @ #130 (NET) BYU. Their other wins are all against awful competition at home. You theorize away games being a major factor, but no that's not it. Houston only has ONE away game and ZERO neutral games. In other words, their schedule sucks big time to date. Their schedule gets a lot better in December, but as of today it has to be one of the worst in the nation.

Why do you think the AAC is doing so much better in NET than RPI? It's because the AAC's SOS is horrible. They hardly participated in the tournaments last week, and they haven't had a conference challenge.

No need to be argumentative here, what I'm saying is actually a good thing for the AAC. They may have a better year due to scheduling a bunch of cupcakes. Don't worry, I'm not the NCAA. I won't change it up. I actually prefer it this way being a SJU fan.

How is it that the NCAA overlooked this? I have no clue.

Well, several AAC teams had tournaments the week before which count just as much..... UCF, UConn, Wichita. Pretty much I think everyone but Houston, ECU had a tournament.... So that goes out the door. And tournaments the week before count exactly the same as tournaments last week....

I do think Houston is somehwat overrated right now- but that after tomorrow will change after they see Oregon.

I think one thing that is gone is the thought you can play a great team and lose, and have it benefit you.....
Providence right now has the #139 SOS. A LOT of that is obviously playing Michigan. The 2 teams they lost to are 8-3. The teams that they beat are 12-13. I think their SOS in the NET reflects a lot more the 12-13 part of the schedule instead of the 8-3 part of the schedule. Part of why they are much worse in the NET than they are the RPI. They're looking a lot more at the SOV than they are the SOS.

AAC has a poor SOS relative to major conferences. That's a fact. I'm not trying to bash the AAC here, I am just using their weak SOS and high NET (vs RPI) to illustrate my point.

Here's proof regarding AAC's SOS so we can end this line of discussion:
http://realtimerpi.com/rpi_conf_Men.html

I'm glad we agree that a team like Houston, with a weak SOS, is overrated by NET.

To your last point, yes that's true. Playing a great team and losing does not help in NET. That is what I've been saying, playing weak teams and winning seems to be a lot more beneficial. For RPI that wasn't the case.

I'm not trying to make this an AAC vs BE thing. I'm just trying to figure out how the rankings work.

If conferences/teams that scheduled weak opponents, like AAC/SJU, ended up benefiting from the new system, wouldn't that be incredibly ironic? It'd go against what we all thought.

Maybe there's something we're all missing, and the numbers will work themselves out as the season goes on. Just seems like way too much of an oversight. I know they must've run the model vs previous years prior to finalizing it.

[stever20]

I'd caution against the real time RPI site- in years past when RPI was a lot more trackable- it wasn't anywhere near accurate.... Always way off.

But even if AAC SOS is 23rd- that's only 3 spots off of what it was last year. And that league got 3 top 6 seeds, and had overall team SOS that could be very good. 13/29/32/45/60/67/72/91/97/112/163/186. Definitely able to get teams in the tourney. Cincy was 45, Wichita 29, and Houston was 72. 2 of the better ones- Temple at 13 and UConn at 32 weren't good enough.

One thing for the AAC that's a plus- after this weekend only have 5 games left for USF OOC and only 2 left for ECU. So especially in the ECU case(USF actually better this year)- their damage is pretty much done now.

And I'm just like you trying to figure out the rankings.....

I think the key with this system if you're playing bad teams is you have to win big. You can't just barely squeak by. That seems to be the biggest x factor. I mean, Houston barely beat UTRGV- and their rating dropped by 17 spots.

What seems strange is the system seems to be where if you are playing a bad team, if you win by enough, it won't hurt you that much.... Considering guys like Calipari were involved, and they while they will play a few weak links(see Wednesday, Monmouth as a good example(speaking of that, how much have they fallen?))- they don't play that many..... It seems pretty counter-intuitive. Will be interesting as we start to see Big Ten conference play how that impacts things.

I'd love to see last years numbers to see frankly Providence. They have a lot of keys that are so differing that it's not funny. In some ways- good- great big wins. Great wins away from home. But then struggled against a lot of bad to mediocre teams.....

[paulxu]

Here are Warren Nolan's RPI calculations. If they were still using them (and granted it is early in the season) the BE would be in 3rd place.
3 pts separate the top 4.
10 pts separate #1 and #10

Conference RPI Rank NC Rec NC WP

Big 12 0.6253 1 52 - 12 0.8125
Big Ten 0.6086 2 75 - 20 0.7895
Big East 0.6018 3 50 - 15 0.7692
ACC 0.5908 4 81 - 21 0.7941
SEC 0.5697 5 61 - 28 0.6854
Pac-12 0.5475 6 52 - 22 0.7027
West Coast 0.5425 7 43 - 24 0.6418
Southern 0.5199 8 32 - 26 0.5517
Missouri Valley 0.5185 9 34 - 26 0.5667
American Athletic 0.5161 10 58 - 25 0.6988

[whiteandblue77]

Here are Warren Nolan's RPI calculations. If they were still using them (and granted it is early in the season) the BE would be in 3rd place.
3 pts separate the top 4.
10 pts separate #1 and #10

Conference RPI Rank NC Rec NC WP

Big 12 0.6253 1 52 - 12 0.8125
Big Ten 0.6086 2 75 - 20 0.7895
Big East 0.6018 3 50 - 15 0.7692
ACC 0.5908 4 81 - 21 0.7941
SEC 0.5697 5 61 - 28 0.6854
Pac-12 0.5475 6 52 - 22 0.7027
West Coast 0.5425 7 43 - 24 0.6418
Southern 0.5199 8 32 - 26 0.5517
Missouri Valley 0.5185 9 34 - 26 0.5667
American Athletic 0.5161 10 58 - 25 0.6988

P10