2021-2022 tracking "NET efficiency" and quadrant wins.

[kiyoat]

So, I thought I'd devote an entire thread to tracking the Coyote WBB team's "net efficiency".  Why?  several reasons:

1. The NET score (and quadrant wins) is likely the single most important factor in the NCAA tournament committee's evaluation for at-large teams and for seeding (despite claims that it is not)  This is similar to the past with the RPI.

2. This team is good enough to bother to pay attention to their NET score, because an at-large bid is possible, and a higher seed is possible.

3. "net efficiency" is the primary difference between the RPI and the NET.  They try to make it seem as though they are totally re-inventing the wheel, but in reality, half of the equation (the "Team Value Index") is basically identical to the RPI.  We are already familiar with that.

4. "net efficiency" is a non-standard, advanced metric, but we do know how it is calculated.  I'm surprised there aren't any sites that track that metric for at least the Men's teams.

5. a standard statistic that approximates it is "scoring margin".  It isn't exactly the same thing, but close enough to be worth looking at.

Here are the formulas:

# Possessions = FGA - OREB + TO + (0.4 x FTA)

FGA = field goal attempts
OREB = offensive rebounds
TO = turnovers
FTA = free throw attempts

Offensive efficiency = points scored / possessions x 100

A number greater than 100 is good.

Deffensive efficiency = points allowed / possessions x 100 (yes this is the same quantity)

a number less than 100 is good.

NET Efficiency = Oeff - Deff

a positive number is good.

[kiyoat]

Quadrants:

Quadrant 1: Home 1-30, ---Neutral 1-50, ---Away 1-75
Quadrant 2: Home 31-75, --Neutral 51-100, -Away 76-135
Quadrant 3: Home 76-160, -Neutral 101-200, Away 135-240
Quadrant 4: Home 161-353, Neutral 201-353, Away 241-353

[kiyoat]

Here's a recent article on the re-vamping of the NET for this year:


www.ncaa.com/news/basketball-men/article/2021-10-03/college-basketballs-net-rankings-explained


If you read it, ignore the trashing on the RPI and pumping up the new NET.  They actually lie in this article about how the old RPI failed to consider home vs away games.  The main point is that they have simplified it to two components:

1. the "TVI" or "team value index", which IMO is basically the RPI under a new name.
2. the "net efficiency"

IMO, one of the reasons they are not specific on just giving the formula, is that they want to be able to tinker with it.  They probably weight those two elements, and weight other elements like home vs away, so they can make adjustments to it every few years. (and to avoid public scrutiny and ridicule)

My big take-away is that all the things that mattered to get a good RPI still matter (SOS, road wins, quadrant wins, winning%), but now scoring efficiency (or scoring margin) also matters.  Also, SOS might matter a bit less than it did with RPI.

[kiyoat]

I hadn't realized that Warren Nolan.com LINK tracks "net efficiency" for both women's and mens teams.  Well, that will make things a bit easier....  I'll just keep updating this post to track our Net Efficiency then.  There are enough games played to start.  One thing to keep in mind is that strength-of-schedule is not part of the net efficiency number.  Just performance.  The other side of the NET formula is like the old RPI.  The net efficiency will increase with league play while the RPI decreases.

Date--------Off eff.--------Def eff.---------Net eff.----

11/26/21__81.4(#250)___90.0(#176)__-8.6(#228)
11/29/21__99.5(#59)___107.1(#333)__-7.6(#232)

[Coyote Fan]

I like the NET efficiency quite a bit. I am more of a believer in not the SOS necessarily but how a team plays the schedule they are given. If a team schedules tough but rarely wins any of the games than who really cares.

[kiyoat]

So, since the official NET rankings have come out, and I have two separate rankings compared to RPI and net efficiency scores, this is what I'm thinking:

Looks like the "RPI-like" portion of the formula might be weighted against the "net efficiency" portion at around a 2.55:1 ratio?  Or 2.6:1.

So around 72% of the formula (between 2/3 and 3/4) is the RPI part, and 28% is the net-efficiency part.

This could be way off base, and the RPI part is obviously not the RPI, it is just similar to the RPI. I'm just looking at what would make sense from a ranking standpoint. I'll keep monitoring it, and see if that holds true.