(01-08-2019 01:30 PM)PT_american Wrote:  

(01-08-2019 01:11 PM)fishpro1098 Wrote:  

(01-08-2019 08:27 AM)vick mike Wrote:  

(01-08-2019 08:22 AM)Tiger1983 Wrote:  

(01-08-2019 08:07 AM)CougarRed Wrote:  Bracket Matrix

Houston 4
Cincy 8
UCF 10

Temple the 5th team out.

http://bracketmatrix.com

If Houston continues holding at #4 or above, it will be interesting to see if the Committee allows the Cougars a #1 seed. Anything less than a #2 seed makes a mockery of their own metrics and anything less than a #1 seed requires rationale explanation.

Depends. How far will Houston drop after losing tomorrow night?

Good point.

How high will Temple climb with a win over #4 in net? or will you drop just because??? This thing seems pretty erratic. I do think to maximize bids the top 3 or 4 teams need to beat Houston to add a quality win to their resumes. Say split with Temple, Cinc and UCF and maybe one other in this group of (SMU, Tulsa or Memphis). So maybe Tulsa at the moment but we will see how it goes.

I am kinda questioning the accuracy of this NET, in particular the Net Efficiency calculations. The formula for calculating the Off Eff seems legit, but the Def Eff formula, in particular the Total Number of Opponents Possessions looks suspect.

Offensive Eff, Total Number of Possessions is: FGA - OReb + Turnovers + (.475 * FTA)

Let's look at this:
1. FGA is shots at the basket so it's legit for possession count
2. Subtracting Offensive Rebounds reduces number of possessions which makes Off Eff go up. That's reasonable.
3. Adding turnovers is odd, as it seems more intuitive it would decrease your number of possessions, but have to keep in mind a higher number of possessions actually reduces the Off Eff so I guess that's why they add it.
4. The more free throws you attempt, the higher your possession count, so you are penalized for getting your points at the free throw line? Seems unreasonable IMO.

Now Defensive Efficiency, Total Number Opponent Possessions: FGA - OR + Opp Turnovers + (.475 * Opp FTA)
1. Same start on # of possessions with FGA.
2. Now, subracting your Opponents O Rebounds increases your DE. That sounds backwards, if your oppenents are getting O Rebs, your defenseive efficiency should go down IMO.
3. Adding opponents turnovers? The more your opponent turns the ball over actually reduces your DE, one would think that would increase your defensive efficiency IMO.
4. More opponent free throws reduces your DE. That sounds reasonable.

Everything else looks reasonable I guess. So, it's a nice graphic, but I'm not sure this is an accurate representation of the formula components.

edit: forgot to add my comments on the Net Efficiency = O - D.

So clearly the formula prefers a higher offensive efficiency number versus deffensive efficiency number. Maybe that's why I have questions about the defensive possessions calculation, as it looks like you want the O to be much larger than D if the desired result is a high Net Efficiency. (Is it desired?)
And what is the meaning of a negative Net Efficiency if the D > O?

you aren't punished by more free throws. Your time with the ball is going to either end with a shot, turnover, or free throws. All count as possessions. it makes sense.

Same with defense... yes you are subrtacting rebounds- but then that'll be balanced out by next shot being taken.
adding turnovers- helps your DE. It adds an empty possession....

(01-08-2019 04:10 PM)stever20 Wrote:  you aren't punished by more free throws. Your time with the ball is going to either end with a shot, turnover, or free throws. All count as possessions. it makes sense.

Same with defense... yes you are subrtacting rebounds- but then that'll be balanced out by next shot being taken.
adding turnovers- helps your DE. It adds an empty possession....

Ok, yes, that makes sense for the number of offensive possessions and I figured that's the reason why the FTA is halved (although 3 free throws skews this a bit unfavorably).

So if a high Net Efficiency is desired, then defensive efficiency metric must be small compared with the offensive efficiency metric, so with that in mind:

Defensive Efficiency, Total Number Opponent Possessions: FGA - OR + Opp Turnovers + (.475 * Opp FTA)
1. Same start on # of possessions with FGA.
2. Subracting your Opponents O Rebounds increases your DE metric, which is opposite of desired, so ok, that is reasonable.
3. Adding opponents turnovers decreases your DE metric, which is desired, so ok, that is reasonable.
4. More opponent free throws increases # of opponent possessions, decreasing your DE metric, which is desired.

ok, I'm satisfied then.

I think UCF will be haunted by that Q4 loss. Recently the committee has seemed to put a lot more attention on bad lasses than on good wins.

(01-08-2019 10:06 AM)Stickboy46 Wrote:  

(01-08-2019 09:27 AM)stever20 Wrote:  team sheets are out now
https://extra.ncaa.org/solutions/rpi/Sta...202019.pdf

a few things different from last year
1- they aren't doing an average this year of the 6 metrics like they did last year.
2- they aren't doing an OOC NET like they would do for the RPI
3- up top is different. Last year they had W-L record, Winning percentage rank, SOS, Opp SOS, RPI broken down both total and OOC. This year it's just W-L record, Road Record and SOS.
4- They no longer are doing the SOS within conference.

some interesting changes.

Looking at Houston's resume, they are going to need some help from teams. LSU, OSU, Memphis, Tulsa are all within 4 of falling to a Quadrant below what they are right now.

Their computer numbers are also drastically worse than anyone else in the Top 10. The NET is definitely helping them out this year.

You don't understand. The rankings don't mean squat until March. Their only value is making me feel better. Therefore, mathematically, everything but the NET ranking is bull****.

(01-08-2019 05:26 PM)splitstud Wrote:  

(01-08-2019 10:06 AM)Stickboy46 Wrote:  

(01-08-2019 09:27 AM)stever20 Wrote:  team sheets are out now
https://extra.ncaa.org/solutions/rpi/Sta...202019.pdf

a few things different from last year
1- they aren't doing an average this year of the 6 metrics like they did last year.
2- they aren't doing an OOC NET like they would do for the RPI
3- up top is different. Last year they had W-L record, Winning percentage rank, SOS, Opp SOS, RPI broken down both total and OOC. This year it's just W-L record, Road Record and SOS.
4- They no longer are doing the SOS within conference.

some interesting changes.

Looking at Houston's resume, they are going to need some help from teams. LSU, OSU, Memphis, Tulsa are all within 4 of falling to a Quadrant below what they are right now.

Their computer numbers are also drastically worse than anyone else in the Top 10. The NET is definitely helping them out this year.

You don't understand. The rankings don't mean squat until March. Their only value is making me feel better. Therefore, mathematically, everything but the NET ranking is bull****.

FWIW....

Quote:The NCAA Men’s Basketball Committee has had helpful metrics it has used over the years, and will continue to use the team sheets, but those will now be sorted by the NCAA Evaluation Tool,” Gavitt said. “As has always been the case, the committee won’t solely focus on metrics to select at-large teams and seed the field. There will always be a subjective element to the tournament selection process, too.”

https://www.ncaa.com/news/basketball-men...anking?amp