2023–24 NET Ratings
- Thread starter SFspidur
- Start date
Also why we complain so much about OOC results. They matter.
Happy at 12-5, would be ecstatic at 14-3 (2 of BC, Wich St, UNI). Not sure where our NET would be but record alone would garner some notice.
Happy at 12-5, would be ecstatic at 14-3 (2 of BC, Wich St, UNI). Not sure where our NET would be but record alone would garner some notice.
I still prefer
19. Dayton
79. St. Joe's
81. Richmond
82. Bona
84. Mason
89. UMass
98. black sheep of the A10
101. Duquesne
123. Loyola
134. Davidson
136. GW
161. blue rams of the A10
175. red rams of the A10
222. SLU
223. La Salle
Spiders get a quad 2 win but move up only one spot. Not very encouraging for the possibility of gaining ground in future games.
If the NET predicts a 1pt win, and you get a 2pt win, why would you expect to move much?
Nobody knows exact calculations. But there r 2 parts. 1 being the Team Value Index (TVI), which is a result-based feature that rewards teams for beating quality opponents, particularly away from home.
So we did well in 50% (?) of the stated metric I think a lot of ppl could reasonably think we’d move up at least 1 whole spot. We stayed at 81 didn’t move at all.
Margin of victory (or however that ends up being reflected in efficiency numbers) is so important. Bona shows how it's done...blew out a middling URI team last night and moved from 82 to 71, and it wasn't even on the road.
I understand trying to look at more than just wins and losses, but I hate that they added the margin or efficiency to the rankings.
I caught the end of an ACC game early in the year against some small program. game was over with about a minute left. other team brings in 5 young players who hadn't played all game including a walk on ... giving them the cool moment to be on that floor. ACC team leaves their studs out there and tacks on like 8 points.
I guess it's acceptable because that's how the rankings work now. but there used to be a little sportsmanship at the end of a game.
I caught the end of an ACC game early in the year against some small program. game was over with about a minute left. other team brings in 5 young players who hadn't played all game including a walk on ... giving them the cool moment to be on that floor. ACC team leaves their studs out there and tacks on like 8 points.
I guess it's acceptable because that's how the rankings work now. but there used to be a little sportsmanship at the end of a game.
We just need to keep winning games. The NET will take care of itself if we do that.
to me the NET is a non issue bc of how OOC went.
But the thing that this discussion highlights is that UR is walking a tightrope in victories. A play here or there and these wins could turn into losses. Spiders need more offense on the road and they need to ice games late at the FT line.
Someone pretty accurately posted that UR would have to hold Dukes scoreless last 4m to win (held to 2 pts IIRC.) That is not a position to be in where you can expect regular victories.
But the thing that this discussion highlights is that UR is walking a tightrope in victories. A play here or there and these wins could turn into losses. Spiders need more offense on the road and they need to ice games late at the FT line.
Someone pretty accurately posted that UR would have to hold Dukes scoreless last 4m to win (held to 2 pts IIRC.) That is not a position to be in where you can expect regular victories.
These are all good points.
I like the concept of NET, though:
1. An away victory is weighted
2. Quality of opponent is accounted for
3. Margin is accounted for
When I built my own system years ago, I included these elements. At the time, margin was not usually included, but I argued for it and built it into my system. I did this mainly because it gives teams who can't schedule top 25 teams everyday a chance to prove their worth by winning easily against lesser opponents. It also says something if you're a top 25 team winning by 15 a game vs 1 a game. And if you're a top 25 team and you win against Buffalo by 3 points, you should probably drop.
To me there's also not a huge difference between winning by 1 vs losing by 1 in terms of your rank. Of course there's credit for winning, but I think it is typically overly credited. When you have 360 teams who don't all play each other, you need others ways to rank them.
To Spidermans point, my system had a cap to it. I don't see any significance in winning by 30 vs 20 say. So, in my system, you're not rewarded extra "points" for that additional margin.
You're going to find what appear to be anomalies in every system, certainly if you look at it from a what did my team do last night standpoint. There's like 100 other game data points to add every night. They all factor into the new ranking. I mean we could win on the road against a similarly ranked opponent by a couple points and theoretically drop a few spots. And that would be fair because of what other teams did that night. It's obviously relative and not a closed system in terms of your one games impact on your rank.
I like the concept of NET, though:
1. An away victory is weighted
2. Quality of opponent is accounted for
3. Margin is accounted for
When I built my own system years ago, I included these elements. At the time, margin was not usually included, but I argued for it and built it into my system. I did this mainly because it gives teams who can't schedule top 25 teams everyday a chance to prove their worth by winning easily against lesser opponents. It also says something if you're a top 25 team winning by 15 a game vs 1 a game. And if you're a top 25 team and you win against Buffalo by 3 points, you should probably drop.
To me there's also not a huge difference between winning by 1 vs losing by 1 in terms of your rank. Of course there's credit for winning, but I think it is typically overly credited. When you have 360 teams who don't all play each other, you need others ways to rank them.
To Spidermans point, my system had a cap to it. I don't see any significance in winning by 30 vs 20 say. So, in my system, you're not rewarded extra "points" for that additional margin.
You're going to find what appear to be anomalies in every system, certainly if you look at it from a what did my team do last night standpoint. There's like 100 other game data points to add every night. They all factor into the new ranking. I mean we could win on the road against a similarly ranked opponent by a couple points and theoretically drop a few spots. And that would be fair because of what other teams did that night. It's obviously relative and not a closed system in terms of your one games impact on your rank.
Last edited:
Correct. I think we can have probably a max of 3 losses in conference to be in realistic convo for an at large. That is where we are at right now.
If we win the next 2, than we get a nationally ranked Dayton at home, this will be our opportunity for us to start getting talked about outside of this message board as a legit threat for an at large.
What I don't like about NET is the quads. I don't know exactly how they use this in the calculations, but I can assure you there is no significant difference between beating a quad 2 team on the road ranked #76 vs beating a quad 1 team on the road ranked #75 by the same margin.
Props to the fellas for this 7 game win streak that makes this even worth discussing today.
And a Dayton win would be a requirement of the 15 A10 wins.
I really should dust it off. It was a lot of fun playing with it.
We, of course, are ranked #1. I forgot to mention the factor I built in for 8 legged mascots!
Yeah last night was perfect example - LC stayed at 123 with their 1 point win over Mass at 89 coming into the night. Bonne soared 11 due to margin/efficiency even though URI is now at 182 .. hmmm
Kenpom and other efficiency metrics are great at spotting outliers over time (ie predicting National Champions). When you get into that fat part of the bell curve they are much less useful at distinguishing the 45th best team from the 75th best.
When a stats person with a mid-major agenda takes that concept on there may be some real dialog about how much efficiency should (or should not) go into bubble determination especially when something like running up the score like Bona did last night unarguably means so little.
When a stats person with a mid-major agenda takes that concept on there may be some real dialog about how much efficiency should (or should not) go into bubble determination especially when something like running up the score like Bona did last night unarguably means so little.
And UMass only fell 3 spots because they should have won per their rank, but they only lost by 1.
Meanwhile, URI fell 21 spots because while they were expected to lose, they weren't expected to get blown out as badly as they did. Perhaps if they lost by 1, they may have even moved up.
Exactly. The question becomes how much weight should margin have. I believe it definitely should have some impact, but how much? That's why to me there should be a cap to how much credit achieved. Continuing to run up the score doesn't prove anything.
There was also another fun one I played with having no meaning whatsoever using the transitive property to rank teams based on who beat who. So in some cases the computer would come up with a long list of results declaring that Team Z was the best team. Something like:
Team Z beat Team W
Team W beat Team S
Team S beat Team T
Team T beat Team D
Team D beat Team F
Team F beat Team P
Team P beat Team N
Team N beat Team A
Therefore, of course, Team Z is better than Team A!!
Last edited:
One more point and then I'll shut up for awhile. 😃
Again, without knowing the exact ins and outs, The NET's efficiency rating I expect is much better than the margin approach I used in the past, and that's simply put because some teams are more defensively minded and thus a 10 point win by a defensive team may be equal to a 20 point win by a more offensive team.
Again, without knowing the exact ins and outs, The NET's efficiency rating I expect is much better than the margin approach I used in the past, and that's simply put because some teams are more defensively minded and thus a 10 point win by a defensive team may be equal to a 20 point win by a more offensive team.
Yes, this is a great point. Adjusting for pace, a 10 point win is not always a 10 point win.
I think about this often when they post NBA stats. A team scores 30% more than a team from 20 years ago.
I’d imagine winning ugly in 50s and barely cracking 60 at a reasonable pace is not doing UR any favors tho.
Thats what its about? Winning or losing compared to predictions? That seems odd. I thought ranking had to do with record related to opponents net ranking with emphasis on home/away ie. Quad games.
If the number 5 team in the country beats the number 350 team should they move up? If you perform as expected based off the current ranking and your opponents ranking then your current ranking is correct and should not change. Seems pretty straightforward to me.
In fact, if the #5 team doesn't beat the #350 by a pretty good margin, they should probably move down.
ok from what I’ve seen Im ready to die on the hill that the NET rewards margin of victory WAY too much. Multiple people have mentioned a 1 pt expected victory on the road being not that different from a 1 point loss. That is totally backwards and a bit of a flaw with the system. A 10 pt victory vs a 1 point victory should barely matter in this ranking. And a huge difference should come from Ws and L’s despite whatver efficiency or expected outcome nonsense
i get the 5 vs 350 comparison yada yada but the specific example of us not moving up much with a win at Duquesne, whereas Bona has a huge jump from their margin of victory beating Rhody, you can’t tell me that’s not backwards. I understand this is impossible with so many variables with 362 teams, but if anything our win over Bona should be the magical factor that moves us ahead of them in the net, not their inconsequential margin vs Rhody.
Some of the other metrics the committee relies essentially only consider margin of victory in their calculations, and don't explicitly consider wins and loses at all. NET is about 50/50. RPI had 0 MOV considerations and it was pretty bad compared to the other metrics by essentially any measure.
Metrics that consider MOV and those that don't are essentially measuring different things about the teams. One is evaluating the record of games that have already occurred, the other is telling you how likely the team is to win future games. These can sometimes give very different pictures. If you are betting, you better be considering MOV.
Makes total sense 2011, appreciate that context. Definitely some funky teams on the RPI list (although Dayton fans may want to become RPI truthers)
100% disagree in terms of a ranking system. There is certainly weight given to a win, but it doesn't determine everything when you are trying to rank 362 teams that don't all play each other.
Aside from the victory points the system automatically awards from winning, there is a VERY slim difference between a 1 point loss and a 1 point win. One play can flip that outcome. There is a much bigger difference between a 1 point victory and a 10 point victory. Multiple plays at least, wouldn't you say, and probably many including made shots and defensive stops.
To make the NCAAs as a mid major a team probably needs to be in the mid to low 40s Net, unless you win your conference. It is a very low probability we get there, but I do think we have a decent shot at top 4 in conference which gives us chance to win the conference. Our NET is important to make the NIT if we don't make the NCAAs.
