Massey #1 | The Boneyard

Massey #1

Status
Not open for further replies.
.-.
Yesterday seemed to be "pick a cupcake to play" day for the top teams. Utah sure seemed a step down after Michigan, but then I saw that South Carolina played a school named Queens University, and rolled 121 points on those poor kids. You don't get to be number 1 after that.
 
Man, I feel like I need to patrol every thread...
homer simpson episode 10 GIF
 
I guess I don’t get it? It looks like SC is #1 on this chart.
 
.-.
Kat said:
Yesterday seemed to be "pick a cupcake to play" day for the top teams. Utah sure seemed a step down after Michigan, but then I saw that South Carolina played a school named Queens University, and rolled 121 points on those poor kids. You don't get to be number 1 after that.
Expand ...
It’s pretty much been, in the words of Dick Vitale, “Cupcake City Baby!” Other than a couple of real basketball games yesterday and UConn’s typically strong OOC schedule, a lot of top teams are playing more cupcakes than you find at a Junior HS bake sale. I think LSU set a record reeling off 6 straight games when the scored 100 pts or more playing against the Sisters of the Poor.

Obviously, the P4 teams will pick up their competition once they start conference play. But at the risk of riling up the TN haters around here, Pat Summit’s philosophy that they would play “anyone, anywhere, anytime” was laudable, but no longer seems relevant for most teams.
 
oldude said:
It’s pretty much been, in the words of Dick Vitale, “Cupcake City Baby!” Other than a couple of real basketball games yesterday and UConn’s typically strong OOC schedule, a lot of top teams are playing more cupcakes than you find at a Junior HS bake sale. I think LSU set a record reeling off 6 straight games when the scored 100 pts or more playing against the Sisters of the Poor.

Obviously, the P4 teams will pick up their competition once they start conference play. But at the risk of riling up the TN haters around here, Pat Summit’s philosophy that they would play “anyone, anywhere, anytime” was laudable, but no longer seems relevant for most teams.
Expand ...
It must not matter analytically in terms of ranking and/or seeding. That could be bad for the product and fan enjoyment. The kids miss out too.
 
Texas could be number one tomorrow. Those were two very impressive wins, particularly with three rotation players injured.
 
Love looking at LSU's SoS.
I'm guessing (?) that the ratings still include some of last season's (tournament) games? Otherwise, wouldn't Texas' SoS be better than ours?
 
.-.
Centerstream said:
Love looking at LSU's SoS.
I'm guessing (?) that the ratings still include some of last season's (tournament) games? Otherwise, wouldn't Texas' SoS be better than ours?
Expand ...
UConn’s opponents Ohio St., Louisville, MI and FSU have SoS as 8, 9, 10 and 12, respectively. Others are Utah (31) and Loyola (72).

TX’s opponents UCLA and South Carolina have SoS as 4 and 6, respectively. Others are Richmond (26), James Madison (32), Incarnate World (134), Louisiana-Lafayette (178), and Texas-Southern (359).

Massey’s undergrad thesis normally solves an n x n system of equations for the Ratings (maximum likelihood Bayesian estimation).

If I am reading this FAQ correctly, the pure Margin of Victory (MOV) model of his thesis (above) is proprietarily adjusted so that the MOV goes through a optimized normalization function that transforms the MOV to a number between 0 and 1 that is then input in the n x n equations. The transformation accounts for blowouts, venue and implicitly for strength of schedule (SoS). The “optimization” solves for the “equilibrium” transformation (per some measure) that yields Ratings and SoS.
 
Centerstream said:
Love looking at LSU's SoS.
I'm guessing (?) that the ratings still include some of last season's (tournament) games? Otherwise, wouldn't Texas' SoS be better than ours?
Expand ...
Massey does indeed start out with last seasons ratings and then diminish their weight. I don’t recall at what point last years value is gone.
 
NycUcWbbFan said:
UConn’s opponents Ohio St., Louisville, MI and FSU have SoS as 8, 9, 10 and 12, respectively. Others are Utah (31) and Loyola (72).

TX’s opponents UCLA and South Carolina have SoS as 4 and 6, respectively. Others are Richmond (26), James Madison (32), Incarnate World (134), Louisiana-Lafayette (178), and Texas-Southern (359).

Massey’s undergrad thesis normally solves an n x n system of equations for the Ratings (maximum likelihood Bayesian estimation).

If I am reading this FAQ correctly, the pure Margin of Victory (MOV) model of his thesis (above) is proprietarily adjusted so that the MOV goes through a optimized normalization function that transforms the MOV to a number between 0 and 1 that is then input in the n x n equations. The transformation accounts for blowouts, venue and implicitly for strength of schedule (SoS). The “optimization” solves for the “equilibrium” transformation (per some measure) that yields Ratings and SoS.
Expand ...
Huh!🤔
 
NycUcWbbFan said:
UConn’s opponents Ohio St., Louisville, MI and FSU have SoS as 8, 9, 10 and 12, respectively. Others are Utah (31) and Loyola (72).

TX’s opponents UCLA and South Carolina have SoS as 4 and 6, respectively. Others are Richmond (26), James Madison (32), Incarnate World (134), Louisiana-Lafayette (178), and Texas-Southern (359).

Massey’s undergrad thesis normally solves an n x n system of equations for the Ratings (maximum likelihood Bayesian estimation).

If I am reading this FAQ correctly, the pure Margin of Victory (MOV) model of his thesis (above) is proprietarily adjusted so that the MOV goes through a optimized normalization function that transforms the MOV to a number between 0 and 1 that is then input in the n x n equations. The transformation accounts for blowouts, venue and implicitly for strength of schedule (SoS). The “optimization” solves for the “equilibrium” transformation (per some measure) that yields Ratings and SoS.
Expand ...
Thank you. Now I understand.
 
As Massey continues to integrate this season’s data, UConn has just surpassed South Carolina in the Power column. This calculation changes more slowly than Rating as SC has been #1 since last season, even after UConn won the championship.

Welcome, Massey, to the 2026 story line…
 
For those who might be wondering- the Massey ratings have nothing to do with NCAA tournament seeding or rankings. (A reminder that AP and Coaches poll rankings also have nothing to do with seeding). Massey is for numbers wonks and maybe some casual gamblers
 
.-.
VaConnFan11 said:
I like Bart Torvik's better (barttorvik.com). Seems to have stability in the rankings. No huge changes from 1 game to the next.
Expand ...
Stability is nice but not at the expense of on court proof. That ranking has undefeated Texas ranked below SCar and UCLA whom the Longhorns have defeated. It's a good site though, and does provide a lot of good data used by the algorithm.
 
FanFromMaine said:
As Massey continues to integrate this season’s data, UConn has just surpassed South Carolina in the Power column. This calculation changes more slowly than Rating as SC has been #1 since last season, even after UConn won the championship.
Expand ...

According to Massey’s site, their final 2025 ratings had UConn ahead of South Carolina, albeit by a very slim margin:

Massey Ratings - College Basketball Women's : NCAA D1 Ratings

Computer ratings and rankings for CBW (College Basketball Women's), with links to team predictions, scores, and schedules.
masseyratings.com masseyratings.com
 
visitingcock said:
Stability is nice but not at the expense of on court proof. That ranking has undefeated Texas ranked below SCar and UCLA whom the Longhorns have defeated. It's a good site though, and does provide a lot of good data used by the algorithm.
Expand ...
Torvik's method doesn't rely heavily on the outcome of one game. I believe his methodology tries to give in-site to what would be the outcome if the teams played 100 times. In the games you mentioned, his system probably "thinks" that SCar and UCLA would still beat Texas maybe 51 times out of 100. Texas just happened to take the 1st actual game.

Luckily, the best teams on paper don't always win. Otherwise it wouldn't be any need to play the games.
 
Other than UCONN , Villanova is the highest ranked Big East team at 45. Ouch.
 
VaConnFan11 said:
I like Bart Torvik's better (barttorvik.com). Seems to have stability in the rankings. No huge changes from 1 game to the next.
Expand ...
I look at most of the statistical sites, and like the one you referenced too, in particular their Plus/Minus rating per 100 possessions, adjusted for level of competition can be viewed for each player game by game, not just for the entire season.

You just click on a specific player and scroll down to their game by game results to see how they performed in each game. at least by their statistical metrics under the bpm column.
 
.-.
Status
Not open for further replies.

Forum statistics

Threads
168,006
Messages
4,549,039
Members
10,431
Latest member
TeganK


Top Bottom