ADVERTISEMENT

2023–24 NET Ratings

Quadrant 1 == 1-5
Quadrant 2 == 4-4
Quadrant 3 == 6-0
Quadrant 4 == 6-0

If that is a #26 ranked NET sheet then I'm the King of Denmark. I know the NET lovers will be along shortly to tell me how far off-base I am, I don't care.
Haven't looked at the NET but it's all about playing as many Q1s as possible, win or lose it's a win. That's why so many of us have been saying schedule more difficult OOC games. Though I know some here say that's very difficult to accomplish. It's almost like a cliche these days ..... the John Chaney way.

Off hand anyone, how many Q1 games on Spider resume?
 
Haven't looked at the NET but it's all about playing as many Q1s as possible, win or lose it's a win. That's why so many of us have been saying schedule more difficult OOC games. Though I know some here say that's very difficult to accomplish. It's almost like a cliche these days ..... the John Chaney way.

Off hand anyone, how many Q1 games on Spider resume?

Q1 games in the A10:

Dayton - 3-3
SBU - 0-4
Spiders - 1-2
vcu - 1-3
UMass - 1-2
GMU - 0-2
SJ - 1-3
LOY - 0-2
Duq - 1-3
Davidson - 0-4
URI - 0-4
Fordham - 1-2
GW - 0-3
SLU - 0-3
LaSalle - 0-3
 
  • Like
Reactions: urfan1
Here is a question for people championing RPI. Do you really think that MWC has 3 of the top 10 teams in the country right now? There are holes in every ranking system, people actually don't agree on what is even being ranked by these systems. There isn't any one perfect ranking that everyone agrees on, because different rankings are developed for different purposes and measure different things. If you understand how the committee uses the NET, and what the NET intends to measure, it makes a lot more sense than RPI.
 
Last edited:
  • Like
Reactions: ChoppinBroccoli
KenPom has them at 39, Torvik at 34, so it's not strictly a NET issue.
 
Q1 games in the A10:

Dayton - 3-3
SBU - 0-4
Spiders - 1-2
vcu - 1-3
UMass - 1-2
GMU - 0-2
SJ - 1-3
LOY - 0-2
Duq - 1-3
Davidson - 0-4
URI - 0-4
Fordham - 1-2
GW - 0-3
SLU - 0-3
LaSalle - 0-3
Now we see why we probably are a one bid league. Depending how things shake out with Dayton, maybe a 2 bid league at best.
 
  • Like
Reactions: 8legs1dream
The maddening thing is that the A10 gets so devalued every year, but how many mid-level Big10, ACC, etc. teams are going into Dayton, UR, vcu, Bonaventure, UMass and Loyola this year and sweeping those games? I bet none. But basically we have to go 16-2 in the league to sniff a bid. I get that our OOC didn't do us any favors, but we've lost ONE home game all season, and after that one (to a team that appears to actually be pretty decent), the consensus was, Well, they lost at home to UMass, so they're done. Meanwhile you have teams like Texas sitting there 6-7 in their league and people probably figure they could afford a few more losses.
 
Take a look at Q1 OOC games, since those are the ones you have some degree of control over. The A-10 was 4–22 with 3 of the wins by Dayton. That obviously doesn't set your conference up for success heading into the phase where you beat each other up.

Dayton: 3–2
Bona: 0–2
UR: 0–2
VCU: 0–2
UMass: 0–0
Mason: 0–1
St. Joe's: 1–1
Loyola: 0–2
Duquesne: 0–1
Davidson: 0–2
URI: 0–2
Fordham: 0–1
GW: 0–1
SLU: 0–2
La Salle: 0–1
 
Here is a question for people championing RPI. Do you really think that MWC has 3 of the top 10 teams in the country right now? There are holes in every ranking system, people actually don't agree on what is even being ranked by these systems. There isn't any one perfect ranking that everyone agrees on, because different rankings are developed for different purposes and measure different things. If you understand how the committee uses the NET, and what the NET intends to measure, it makes a lot more sense than RPI.

Agreed none perfect all have flaws. But correct me if I’m wrong, wasn’t the RPI formula known? It was used for a long time until teams got more familiar w it and non p6 teams especially performed better. Then they changed to a secret formula that seemingly awards p6 teams higher rankings. Call that cynical if u want. We don’t know the NET formula so I think it’s hard to say it makes more sense. I think Mooney has even spoken to this subject. U don’t know what u need to do. If the formula is known like RPI was well at least u can target what it wants. For instance the margin/efficiency stuff. And you’d probably find non p6 start to do better again. Tho there is less inventory of available OOC games now.
 
  • Like
Reactions: urfan1
That's why so many of us have been saying schedule more difficult OOC games. Though I know some here say that's very difficult to accomplish. It's almost like a cliche these days ..... the John Chaney way.
things have changes a lot since the John Chaney days.

there was a recent article about Princeton. at 17-3, they're 53 in the NET and 22 in RPI. they're 0-0 in Q1 games and thus are not in the running for an at-large despite making the Sweet 16 last year. their claim ... "we called everybody. we played the best teams that would play us".

they say they weren't even asking for home and homes. they knew nobody would come to them. but teams wouldn't even play them on their own courts. Princeton's a tough team to prep for and they're good. there's not much upside.

Rutgers agreed to host after I think the Governor got involved, lol. "The Battle of New Jersey". Princeton won.

there's no answer. you can't force high majors to play mid majors. and you certainly can't make them go on the road. maybe offer bonus NET points for a P6 to play a non-P6 on the road? but then you have to exclude teams like Gonzaga and Memphis.
 
Agreed none perfect all have flaws. But correct me if I’m wrong, wasn’t the RPI formula known? It was used for a long time until teams got more familiar w it and non p6 teams especially performed better. Then they changed to a secret formula that seemingly awards p6 teams higher rankings. Call that cynical if u want. We don’t know the NET formula so I think it’s hard to say it makes more sense. I think Mooney has even spoken to this subject. U don’t know what u need to do. If the formula is known like RPI was well at least u can target what it wants. For instance the margin/efficiency stuff. And you’d probably find non p6 start to do better again. Tho there is less inventory of available OOC games now.
RPI formula was secret for the first ~20 years it was used.
 
Well u could force them. Go to some type of professional scheduling model. Equal amount of home away games. Problem is schools no longer have any respect for a regulatory authority like ncaa. But if we r in the era of amateur athletics now being professional then treat scheduling the same too.
 
  • Like
Reactions: dobbiethespider
RPI formula was secret for the first ~20 years it was used.

Ok thx & how long was it then known? Maybe their intention is to make NET known eventually too but I’m not sure what argument is to be secret. But again I think they want it secret bc it’s helping p6. The RPI was fine when secret until it wasn’t after no longer secret and non p6 improved. I admittedly haven’t done a deep dive but that’s how it feels from a macro level.
 
And look I like metrics. U have to have them. Not an “eye test” guy that just favors big teams more. Just a little weird to me when I hear proponents of NET say it’s better. Maybe, but it’s secret how do u know. We have this new #1 measurement tool to rank teams, we can’t tell u how it’s done but trust us it’s AWESOME. In caps.
 
Well u could force them. Go to some type of professional scheduling model. Equal amount of home away games. Problem is schools no longer have any respect for a regulatory authority like ncaa. But if we r in the era of amateur athletics now being professional then treat scheduling the same too.
Guess you could try to make them, but then you've got to make us too. I don't see us playing at Radford or Longwood too often.
 
I get the criticisms of RPI...75% of your rating is just how your opponents and your opponents' opponents perform. Other than creating a schedule prior to the season, you have no control over 75% of your rating. And among the 25% you do control, there's no margin of victory or efficiency factor.

It's a very simplistic system that actually worked remarkably well for how simple it was.

NET was supposed to be "better" due to its ability to incorporate additional metrics beyond just wins and losses. Whether it is better or not depends on your point of view.
 
  • Like
Reactions: gospidersgo
RPI formula was secret for the first ~20 years it was used.
Really that long?

When did it start, when did Jerry Palm seem to reverse engineer it? And when did it start using location?

I feel like there we 3 periods of RPI, unknown, known, and location weighted.
 
Good article on the history of the RPI through 2003. Debuted in 1981 and looks like it was being reverse engineered about 10 years later. Was tweaked a couple of times over the years, and in fact it originally had a road factor included but it was quickly removed.

 
  • Like
Reactions: fan2011 and urmite
Good article on the history of the RPI through 2003. Debuted in 1981 and looks like it was being reverse engineered about 10 years later. Was tweaked a couple of times over the years, and in fact it originally had a road factor included but it was quickly removed.

I feel I started following it about Beilein‘s first season, or maybe 1996. Either way after the 1993 change. And remember Palm at-large predictions becoming less reliable after location adjustment was added.
 
Good article on the history of the RPI through 2003. Debuted in 1981 and looks like it was being reverse engineered about 10 years later. Was tweaked a couple of times over the years, and in fact it originally had a road factor included but it was quickly removed.

I believe RPI formula wasn’t actually released until after that article was published, in 2006. Before then the RPI also included mysterious arbitrary adjustments beyond the factors that are known today, and these adjustments were never fully reverse engineered. The article mentions them and how the RPI publishers at the time don’t get it fully right.
 
UR: 71-> 69
Bonne: 69->79 (causing our win to move to Q3)
VCU: 78 -> 74 (so our loss goes back to Q1)
Dayton stays at 19 after their fabulous loss against all powerful GMU.
 
  • Haha
Reactions: Eight Legger
UR: 71-> 69
Bonne: 69->79 (causing our win to move to Q3)
VCU: 78 -> 74 (so our loss goes back to Q1)
Dayton stays at 19 after their fabulous loss against all powerful GMU.
Was that Dayton’s first non Q1 loss?
 
All this makes my head hurt. The quad thing confuses me. In the final quad rankings, is it what the opponent is ranked when we play them, or is it what they are at the end of the season?
If the answer is "end of the season", then why even have these discussions at all right now???
 
  • Haha
Reactions: PalmTreeSpidUR
Does anyone know, what is the reason for having the formula be a secret? I like transparency 🤷🏼‍♂️
 
To prevent teams from gaming it, like they did for the RPI. See Goodharts Law.

Goodhart is probably a good fellow, a right jolly old elf. But I bet he knows zip about sports and it wasn't intended to be used for sports. Teams want to know how to best qualify for post season, justifiably so.

GKiller's Law: NCAA Tournament is controlled primarily by P6, we have a measurement formula that rewards P6 teams, let's keep it a secret as long as it continues to help P6 and then change it again if it starts to hurt.
 
GKiller's Law: NCAA Tournament is controlled primarily by P6, we have a measurement formula that rewards P6 teams, let's keep it a secret as long as it continues to help P6 and then change it again if it starts to hurt.
Bingo
 
Goodhart is probably a good fellow, a right jolly old elf. But I bet he knows zip about sports and it wasn't intended to be used for sports. Teams want to know how to best qualify for post season, justifiably so.

GKiller's Law: NCAA Tournament is controlled primarily by P6, we have a measurement formula that rewards P6 teams, let's keep it a secret as long as it continues to help P6 and then change it again if it starts to hurt.
Goodhart's law isn't about sports, it was originally stated about economics but has been shown to apply generally. Metrics stop being useful when they become a target. This has been shown over and over again, and is common sense really. Metrics all have flaws, a single number can't capture the nuance of complicated systems. When a metric becomes a target (e.g. I want to get the best RPI/NET to maximize my post season chances) people will naturally take advantage of its flaws and it ceases to be as useful as it was before, when people were not trying to game the metric.
 
Last edited:
  • Like
Reactions: PalmTreeSpidUR
But when a system is biased to benefit a portion of the population over another portion, and the unbiased portion finds a way to "game" the system to remove some of the built in bias? Is that truly gaming the system?
 
But when a system is biased to benefit a portion of the population over another portion, and the unbiased portion finds a way to "game" the system to remove some of the built in bias? Is that truly gaming the system?
Why would you assume only the "unbiased portion" would game the system? The "biased portion" usually has much more power to game the system.
 
Palm slides us back in as the auto, noting we're a bid stealer who otherwise wouldn't be in the field. Dayton sitting at a 6-seed.



Lunardi hasn't updated today.
Palm states the A-10 is one bid league if Dayton wins the auto bid as well. So, yeah, still an uphill battle for us. Keep stacking wins.
 
  • Like
Reactions: urfan1
Yeah, I mean, maybe if we win out until A-10 final we'd have a shot, but as it stands right now, Dayton is clearly the only at-large candidate.
 
  • Like
Reactions: urfan1
Yeah, I mean, maybe if we win out until A-10 final we'd have a shot, but as it stands right now, Dayton is clearly the only at-large candidate.
Think a 16-2 A-10 team would be tough to leave out, but sadly it is a real uphill battle. We didn't make enough hay in the OOC.
 
  • Like
Reactions: urfan1
ADVERTISEMENT

Latest posts

ADVERTISEMENT