NET Rankings / Bracketology

Status
Not open for further replies.
#51
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.
I agree, and similarly those 40 teams will continue to lose games. So keep on winning and the ranking will come.
 
#53
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.
That's because models like KenPom and Torvik don't look at wins and losses, they look at offensive and defensive efficiency relative to the opponent's offensive and defensive efficiency. A team could lose every game on their schedule but would likely be ranked #2 if each of those games was a one point loss to the #1 team.

It's also worth noting that the rankings aren't really rankings in the traditional sense. They're just an ordering of all teams based on their efficiency margin (KenPom) or expected win percentage vs. an average team (Torvik). For example, the difference between the #20 and #24 ranked teams on KenPom right now is less than 1/10 of a point. On the other hand, the difference between #1 and #5 is five points. It's generally more informative to pay attention to the numbers rather than the ranking.
 
Likes: RML
#55
st petersburg, fl
Some love from ESPN, Well...I guess you could consider being called an overachiever love. :unsure:

Overachieving: Illinois. With Kofi Cockburn on the floor, Illinois is one of the best teams in the country. In those situations, Illinois has connected on 53.4% of its attempts inside the arc and held opponents to just 87 points per 100 possessions. Per the Chicago Tribune, a preseason Big Ten media poll projected Brad Underwood's squad to finish seventh. Instead, the program entered the week tied with Michigan State for first place (both 7-2) and chasing a top-eight seed in the NCAA tournament. Saturday's thrilling road win over Michigan was its sixth in a row and third Big Ten road win.

https://www.espn.com/mens-college-b...erachievers-underachievers-top-leagues#BigTen

And, related (scrolls further down)

19. Illinois Fighting Illini

Three of the past five champions have ranked as a top-30 team in rebound rate AND top 70 in field goal percentage. Thus far, the Illini rank fourth and 45th, respectively. The Illini are a top-10 team in terms of minute continuity, according to KenPom data, and they are will to battle with anyone. Don't sleep on them!
 
#56
Some love from ESPN, Well...I guess you could consider being called an overachiever love. :unsure:

Overachieving: Illinois. With Kofi Cockburn on the floor, Illinois is one of the best teams in the country. In those situations, Illinois has connected on 53.4% of its attempts inside the arc and held opponents to just 87 points per 100 possessions. Per the Chicago Tribune, a preseason Big Ten media poll projected Brad Underwood's squad to finish seventh. Instead, the program entered the week tied with Michigan State for first place (both 7-2) and chasing a top-eight seed in the NCAA tournament. Saturday's thrilling road win over Michigan was its sixth in a row and third Big Ten road win.

https://www.espn.com/mens-college-b...erachievers-underachievers-top-leagues#BigTen

And, related (scrolls further down)

19. Illinois Fighting Illini

Three of the past five champions have ranked as a top-30 team in rebound rate AND top 70 in field goal percentage. Thus far, the Illini rank fourth and 45th, respectively. The Illini are a top-10 team in terms of minute continuity, according to KenPom data, and they are will to battle with anyone. Don't sleep on them!
I think overachiever is a fair term in our case. I think many figured that this team would finish anywhere from 5th to 9th....right there on the bubble. After the Missouri loss, the expectations dropped to 11th or 12th with some wanting the coach fired (per usual here).

Well, a month later....6 straight wins, three at very difficult road venues and tied for 1st. I'd say that currently, we're surpassing expectations amongst those not wearing O & B glasses or drinking the kool-aid.
 
#57
That's because models like KenPom and Torvik don't look at wins and losses, they look at offensive and defensive efficiency relative to the opponent's offensive and defensive efficiency. A team could lose every game on their schedule but would likely be ranked #2 if each of those games was a one point loss to the #1 team.

It's also worth noting that the rankings aren't really rankings in the traditional sense. They're just an ordering of all teams based on their efficiency margin (KenPom) or expected win percentage vs. an average team (Torvik). For example, the difference between the #20 and #24 ranked teams on KenPom right now is less than 1/10 of a point. On the other hand, the difference between #1 and #5 is five points. It's generally more informative to pay attention to the numbers rather than the ranking.
Oh, I get that. that's why I made a comment about not being the most efficient team.

On a related (computer ranking) note....didn't most of those computer models start capping the Margin of Victory factor at 10 last year? The fact we've had quite a few close games (@Michigan, v NW, @Wisc, v RU) probably factors in to a lower ranking along with our offensive inefficiency. Beating Purdue by 26 is the same as beating them by 10 if Margin has truly been capped.
 
#58
I think overachiever is a fair term in our case. I think many figured that this team would finish anywhere from 5th to 9th....right there on the bubble. After the Missouri loss, the expectations dropped to 11th or 12th with some wanting the coach fired (per usual here).

Well, a month later....6 straight wins, three at very difficult road venues and tied for 1st. I'd say that currently, we're surpassing expectations amongst those not wearing O & B glasses or drinking the kool-aid.
Experts undervalued Kofi’s affect on the team as well as BU’s adaptation. Said before the season they had talent to match up with anyone wasn’t sure coaching was as good as recruiting. Think they have proven it is.
 
Likes: the national
#62
M tipping over
Pdx
I think overachiever is a fair term in our case. I think many figured that this team would finish anywhere from 5th to 9th....right there on the bubble. After the Missouri loss, the expectations dropped to 11th or 12th with some wanting the coach fired (per usual here).

Well, a month later....6 straight wins, three at very difficult road venues and tied for 1st. I'd say that currently, we're surpassing expectations amongst those not wearing O & B glasses or drinking the kool-aid.
I think if BU didn't change something about the way we were playing we would have probably wound up where most people forecasted us to be.
 
#66
Item #1 looks to be the only black box. Seems like with only 3 variables unknown and nearly 2 years of data it should be possible to back calculate some factors to come close to determining the actually equation. The hardest part would be determine how “opponent” is factored in.
To illustrate, we didn't play yesterday, and neither did Wisconsin. But other teams did, and as a result we swapped places with Wisconsin in the NET rankings last night.:)
 
#67
New York
1 looks like it might be fairly close to RPI
 
#68
The Transfer Portal
It would be helpful to get a top 4 seed to ensure that. Otherwise it will take some luck.
Two things on that:

1) I feel like every year there’s at least one “protected” seed that gets a tough break in that they are technically close to home, but some other lower seed is even closer in their pod. Would love to be the object of someone’s ire who (rightfully) complains that “protection” apparently means getting Illinois in St. Louis.

2) If the B1G really gets like 10 teams in, it’ll add more restrictions on where teams can go. We could get the “lower seed/better location” combo the committee sometimes hands out because there will be some regions/pods we just can’t be placed in because of other B1G teams.
 
#69

Deleted member 17568

D
Guest
Two things on that:

1) I feel like every year there’s at least one “protected” seed that gets a tough break in that they are technically close to home, but some other lower seed is even closer in their pod. Would love to be the object of someone’s ire who (rightfully) complains that “protection” apparently means getting Illinois in St. Louis.

2) If the B1G really gets like 10 teams in, it’ll add more restrictions on where teams can go. We could get the “lower seed/better location” combo the committee sometimes hands out because there will be some regions/pods we just can’t be placed in because of other B1G teams.
Regarding #2, it is highly unlikely the B1G gets more than 4 seeds in the top 4. They would never drop a protected seed to an unprotected to seed. They would drop a 5 to a 6 instead, for example.
 
#70
The Transfer Portal
Regarding #2, it is highly unlikely the B1G gets more than 4 seeds in the top 4. They would never drop a protected seed to an unprotected to seed. They would drop a 5 to a 6 instead, for example.
Agree - I’m more referring to if we play someone we’ve already played twice in the BTT (Maryland, for instance), then we can’t be in their region at all.
 
#71
Item #1 looks to be the only black box. Seems like with only 3 variables unknown and nearly 2 years of data it should be possible to back calculate some factors to come close to determining the actually equation. The hardest part would be determine how “opponent” is factored in.
Black boxes are OK. OK, they are not.

The issue, when it comes to such an "algorithm" (in a black box (best of my knowledge this algo is an unknown equity)) is that there is really no recourse to lodge a protest.


So, I for one will wait till the end of the season to look at the NET (= brackets). IMVVHO no use to get one's BP on a consistent slow boil (even in sleep).
 
#72

Deleted member 17568

D
Guest
Agree - I’m more referring to if we play someone we’ve already played twice in the BTT (Maryland, for instance), then we can’t be in their region at all.
I think we can meet in the Elite 8? So if they got a 3 and we got a 4, I believe we could be.

I may be wrong on that though.
 
#73
The Transfer Portal
Agree - I’m more referring to if we play someone we’ve already played twice in the BTT (Maryland, for instance), then we can’t be in their region at all.
I checked on this to make sure I'm not just blindly inventing complications. Let's pretend the season ended today and the BTT seeded out based on the standings.

As the #2, we'd face the winner of Purdue-Penn State. Purdue isn't currently projected in and that'd only be our second game against PSU. We could theoretically then face Rutgers, MSU, or Maryland for the third time if we were to advance to the title game.

I'll just use one bracket to illustrate what could happen if we were to 1) hypothetically beat PSU, 2) beat Rutgers and then 3) lose to Maryland in the BTT title game. Jerry Palm currently has Maryland as a 3, MSU as a 4, PSU as a 4, us as a 5, Rutgers as a 6. We're currently in Maryland's bracket (the East), and if that hypothetical happened we'd have to move (because of the 3 games against the Terps), but:
-We couldn't keep our seed and move to the South, because Penn State is the 4 there and you can't play a team you've already played twice in the second round
-We couldn't keep our seed and move to the West, because MSU is the 4 there and the same thing applies
-We couldn't move to the Midwest at all, because Rutgers is there and we've played them three times already.

So in this scenario, our seed would have to change (or somebody else's would) solely because there's nowhere for us to go as a 5-seed.
 
Likes: the national
#74

Deleted member 17568

D
Guest
I checked on this to make sure I'm not just blindly inventing complications. Let's pretend the season ended today and the BTT seeded out based on the standings.

As the #2, we'd face the winner of Purdue-Penn State. Purdue isn't currently projected in and that'd only be our second game against PSU. We could theoretically then face Rutgers, MSU, or Maryland for the third time if we were to advance to the title game.

I'll just use one bracket to illustrate what could happen if we were to 1) hypothetically beat PSU, 2) beat Rutgers and then 3) lose to Maryland in the BTT title game. Jerry Palm currently has Maryland as a 3, MSU as a 4, PSU as a 4, us as a 5, Rutgers as a 6. We're currently in Maryland's bracket (the East), and if that hypothetical happened we'd have to move (because of the 3 games against the Terps), but:
-We couldn't keep our seed and move to the South, because Penn State is the 4 there and you can't play a team you've already played twice in the second round
-We couldn't keep our seed and move to the West, because MSU is the 4 there and the same thing applies
-We couldn't move to the Midwest at all, because Rutgers is there and we've played them three times already.

So in this scenario, our seed would have to change (or somebody else's would) solely because there's nowhere for us to go as a 5-seed.
We could move to the Midwest. We can play Rutgers in the Elite 8. From here: https://www.ncaa.com/news/basketball-men/article/2018-10-19/how-field-68-teams-picked-march-madness

Each of the first four teams selected from a conference shall be placed in different regions if they are seeded on the first four lines.

Teams from the same conference shall not meet prior to the regional final if they played each other three or more times during the regular season and conference tournament.

Teams from the same conference shall not meet prior to the regional semifinals if they played each other twice during the regular season and conference tournament.

Teams from the same conference may play each other as early as the second round if they played no more than once during the regular season and conference tournament.
 
Last edited by a moderator:
Likes: oNthErILL
#75
I’m under the impression that the NET isn’t a fixed function like RPI that can be calculated explicitly. I think it’s a normalization that is refactored every night. I suspect this is the reason why it isn’t published until the 8th week; they need to accumulate enough data to model it, and determine some sort of normalization constant.

I think that also helps to explain the lack of recently bias, and unpredictable movements in rankings; everyday it’s remodeling the data.
 
Status
Not open for further replies.