NET Rankings / Bracketology
- Status
all the NET Rankings (& bracketology) talk into this thread. Go IlliniIllwinsagain
- Cary, IL
I know that this will change continually for the next 6 weeks or so, but does it seem a little too much of a coincidence to be in the bracket with KU?
sacraig
- The desert
Nah. Just his estimate of how the seeding works out.
cuillini
- San Bernardino, Ca.
Uh oh, the famous 5 - 12 potential upset.......
I think so. Net is an organizational tool, not a seeding tool. I don't think anyone on the committee would even say it's the best tool.
Here's our team sheet:
Illinois
No Q4 losses (worst of the worst)
1 Q3 (bad loss)
1 Q2 (bad, but not as bad as Q3. A lot of teams have Q3 or Q2 losses. Kentucky has a Q4, and 2 Q2 losses, but is 21st in NET. Duke has a Q3 and a Q2, and is 6th)
4-3 vs Q1, which is quite good. Not elite, but if you're .500 or above on Q1, you're solidly in.
My guess is we're a 5 or 6 seed today. Definitely worked our way out of a hole with this winning streak.
Trying to make sense of the NET right now is like trying to put a puzzle together with half the pieces.
It’s an algorithm (or set of algorithms?) that take in specific inputs and provides outputs. It isn’t going to technically be accurate until it has the inputs from every game in the regular season plus conference tourneys. (In regards to using it to discuss NCAA tourney implications)
If we use it to discuss this snapshot in time, we have to take into account our bad losses are 2 games out of 20 as opposed to 2 games out of 30+. So they affect us more heavily right now.
We also have to realize that if we beat a team that is our NET peer, the head to head matchup won’t affect us as much as we’d like. Even though we beat Purdue on the road, it was also a home loss for them and put them at like 10-8 on the season. (Or something like that) The same happened to Michigan. For head to head matchups to have a major affect, it’d have to be something like us beating NET #3 on the road or lose to NET #200 at home.
The real progress we see will be when the teams we beat, continue to win. Unfortunately, the mid level of the B1G is mired with relatively even teams and nobody is pulling away. Also, Miami and Mizzou continue to lose. Miami in particular is a cinderblock tied around our feet.
Our NET will take care of itself when we start to see the teams we beat get to 20+ wins.
A good measure of how our two road wins affected our potential seeding would be to look at the bracket matrix and some of the better bracketologists. We were around a 9 seed a week ago, and are hovering around 6 now.
NET will catch up to that as games are played and if we keep winning.
It’s an algorithm (or set of algorithms?) that take in specific inputs and provides outputs. It isn’t going to technically be accurate until it has the inputs from every game in the regular season plus conference tourneys. (In regards to using it to discuss NCAA tourney implications)
If we use it to discuss this snapshot in time, we have to take into account our bad losses are 2 games out of 20 as opposed to 2 games out of 30+. So they affect us more heavily right now.
We also have to realize that if we beat a team that is our NET peer, the head to head matchup won’t affect us as much as we’d like. Even though we beat Purdue on the road, it was also a home loss for them and put them at like 10-8 on the season. (Or something like that) The same happened to Michigan. For head to head matchups to have a major affect, it’d have to be something like us beating NET #3 on the road or lose to NET #200 at home.
The real progress we see will be when the teams we beat, continue to win. Unfortunately, the mid level of the B1G is mired with relatively even teams and nobody is pulling away. Also, Miami and Mizzou continue to lose. Miami in particular is a cinderblock tied around our feet.
Our NET will take care of itself when we start to see the teams we beat get to 20+ wins.
A good measure of how our two road wins affected our potential seeding would be to look at the bracket matrix and some of the better bracketologists. We were around a 9 seed a week ago, and are hovering around 6 now.
NET will catch up to that as games are played and if we keep winning.
But, but, but, Nicholls State is in second place in their league (half game behind SFA) so that should erase the Miami and Missouri losses /s. Actually, Nicholls State seems to be doing a lot better than originally projected in their league. I think I read they were expected to finish near the bottom.
Seriously, though, seems like the NET powers that be could adopt an outlier rule, take out the worst game that affects any team.
Seriously, though, seems like the NET powers that be could adopt an outlier rule, take out the worst game that affects any team.
The Dave
- CA
What's REALLY stunning to me is that KenPom had us at #21 this morning and Purdue at #20. That's the same Purdue we've beaten at home and ON THE ROAD by 26 and 17. The computer models really don't like us very much. I realize we aren't the most efficient team in the world, but I guess if we keep on playing like we are and winning and a good rate it will all come out in the wash by season's end.
I think this is because Purdue has won by large margins against good teams. Illini only have 1 of those wins against Purdue coincidentally.
Illwinsagain
- Cary, IL
So, next year, run up the score, play and beat teams that only lose to us, don't lose any. That should get us in the top 30! /s
Deleted member 17568
D
Guest
It's the same reason OSU is still so high.
If you are going to do that, you should also take out the best game that affects any team. I suspect we would benefit from doing that.
A few things to keep in mind for the posters just getting into the NET metric...
Q1 = tourney teams
Q2 = NIT teams
Q3 = bottom teams in good conferences and top teams in bad conferences
Q4 = just bad teams
NET 1-30 = guaranteed tourney appearance (only one team ever in NET or RPI top 30 has not made the tourney)
NET 31-45 = 70+% tourney appearance (a few teams per year in top 45 won’t make it for various reasons)(90+% of teams in top 45 will make it)
NET 45-70 = true bubble teams (a few teams will make it from this range and NET ranking itself seems to mean less than other factors such as SOS)
A few discussion points:
- offensive efficiency is a factor
- scoring margin capped at 10 points is a factor
- it uses an adjusted winning precent where a neutral site win = 1 win, a home win = 0.6 wins, and a road win = 1.4 wins
- no recency bias
- its relative in that it scores teams against the field of all teams and not against a fixed standard
- NET is official but committee clearly still references RPI based on past results where teams with very poor NET but good RPI have been selected
Q1 = tourney teams
Q2 = NIT teams
Q3 = bottom teams in good conferences and top teams in bad conferences
Q4 = just bad teams
NET 1-30 = guaranteed tourney appearance (only one team ever in NET or RPI top 30 has not made the tourney)
NET 31-45 = 70+% tourney appearance (a few teams per year in top 45 won’t make it for various reasons)(90+% of teams in top 45 will make it)
NET 45-70 = true bubble teams (a few teams will make it from this range and NET ranking itself seems to mean less than other factors such as SOS)
A few discussion points:
- offensive efficiency is a factor
- scoring margin capped at 10 points is a factor
- it uses an adjusted winning precent where a neutral site win = 1 win, a home win = 0.6 wins, and a road win = 1.4 wins
- no recency bias
- its relative in that it scores teams against the field of all teams and not against a fixed standard
- NET is official but committee clearly still references RPI based on past results where teams with very poor NET but good RPI have been selected
Deleted member 17568
D
Guest
Not to nitpick but Q1 does not necessarily equate to a tournament team. If you beat a top 50 team on a neutral court or a top 75 team on the road, that also counts as a Q1 win.
If NET rankings stay the same as they are today (which they won't), here's how our remaining schedule looks as far quad wins/losses go:
Minnesota - Q2
@Iowa - Q1
Maryland - Q1
MSU - Q1
@Rutgers - Q1
@PSU - Q1
Nebraska - Q3, but bordering on being a Q4.
@NW - Q3
Indiana - Q2
@OSU - Q1
Iowa - Q1
Minnesota - Q2
@Iowa - Q1
Maryland - Q1
MSU - Q1
@Rutgers - Q1
@PSU - Q1
Nebraska - Q3, but bordering on being a Q4.
@NW - Q3
Indiana - Q2
@OSU - Q1
Iowa - Q1
We beat Purdue by large margins twice.
UNC is nolonger a good team. They are last or near ;last in the ACC.
Deleted member 17568
D
Guest
While true, UNC is still better than any of our non-conference wins. OSU beat Villanova by 25, Penn State by 32 and Kentucky on a neutral floor. Their average margin of victory in their 10 non-conference wins was 24.2 points.
I think our NET is pretty well baked in at this point if we finish out between 4-7 and 7-4 as we are playing mostly teams ranked about where we are. We could move up or down maybe 10-15 places. However, the difference in NCAA seeding and the polls between 11-9 and 14-6 in the B1G would be significant.
Basically, the metrics do not favor us. In addition to torvik, I look as David Wilson's rankings at talismanred.com. He has us tied for 33rd overall, and 7th in the B1G. Looking at torvik, when the ratings are viewed for conference games, we are predictably 2nd in the B1G-our barthag score is .8795 overall for 26th, and .9385 in conference for a close second to MSU. Unfortunately, we are being evaluated on the year as a whole.
Basically, the metrics do not favor us. In addition to torvik, I look as David Wilson's rankings at talismanred.com. He has us tied for 33rd overall, and 7th in the B1G. Looking at torvik, when the ratings are viewed for conference games, we are predictably 2nd in the B1G-our barthag score is .8795 overall for 26th, and .9385 in conference for a close second to MSU. Unfortunately, we are being evaluated on the year as a whole.
chrisRunner7
- Spokane, WA
So, I'm not a fan of the NCAA's Net Rankings website... mainly because it doesn't list each team's record versus Q1, Q2, etc., or at least I can't find that info on their site.
There may be better sites, but I discovered BracketResearch.com yesterday. It's got a sortable table that lets you see a better breakdown of where the good wins and bad losses are coming from.
https://bracketresearch.com/team-quadrant-wins-and-losses-tracker/
There may be better sites, but I discovered BracketResearch.com yesterday. It's got a sortable table that lets you see a better breakdown of where the good wins and bad losses are coming from.
https://bracketresearch.com/team-quadrant-wins-and-losses-tracker/
Good point.
I stand corrected, and how can I forgot? Those were great games.
I think the biggest thing people need to understand about advanced stats rankings is there are at least 40-50 other teams that can make a very similar argument to this one. Which is exactly why the stats are needed when ranking teams that play completely different schedules. You can't cherry pick whatever games you please and forget about the rest.
Well luckily it's not. But it is important to take every game into account. We like to think our team has "turned a corner" and will never look back. But we don't know that. Who knows, maybe Ayo and Giorgi start missing shots again. Trent can't find open looks. We start getting antsy on offense and turn the ball over and lose 7 of the next 8.
Ohio St looks terrible right now, but you can't overlook the fact that the team that was on fire early in the season could show up again. That's why they are still so high and we aren't shooting up the rankings like crazy.
Ohio State doesn’t have any bad losses (unlike us) and that’s the main reason.
Of course they also played a really tough non-conference schedule against teams like #13 Kentucky (W), #12 West Virginia (L), and #8 Villanova (W).
Last edited:
- Status