Good Ole Massey rankings

[visitingcock]

Oregon Jumps to Number 2. UConn falls to 4. NC State falls 10 spots. Texas rises 10 spots.. Massey Ratings - College Basketball Women's : NCAA D1 Ratings

Stanford Pac 12	0-0 0.000		1 9.32	1 69.34	4 100.02	3 33.99	2.90	2 69.34	1 55.42	22.94	6.06
Oregon Pac 12	1-0 1.000	+4	2 9.15	2 67.78	10 97.77	1 34.67	2.99	171 36.82	9 52.88	21.01	5.99
South Carolina Southeastern	1-0 1.000	+2	3 9.09	4 66.02	12 97.71	5 32.97	2.90	14 61.48	6 53.12	20.40	5.60
Connecticut Big East	0-0 0.000	-2	4 9.02	3 66.69	5 99.30	6 32.05	2.90	5 66.69	25 51.17	23.70	5.30
Maryland Big 10	1-0 1.000	-1	5 8.88	5 65.42	1 104.87	47 25.22	2.87	252 27.02	16 52.09	21.36	7.64

 

[Frodo]

Massey ratings will get sorted out pretty quickly, unlike RPI, which will still be nonsense a couple of months from now…thank god the ncaa finally got a clue and dumped RPI in favor of the NET. I still wish they had just opted to use Massey or Sonny Moore

 

[Plebe]

You're really perseverant on hyping Massey when it's the least relevant.

Massey is virtually meaningless this early in the season. Give it at least 8-10 games to gather current season data please.

 

[Boreifs]

You're really perseverant on hyping Massey when it's the least relevant.

Massey is virtually meaningless this early in the season. Give it at least 8-10 games to gather current season data please.

I don’t even look until January

 

[Kaizen]

Massey rankings are purely based on computer. If Stanford and UConn haven't play any games how can computer rank them?

It is a joke.

 

[visitingcock]

You're really perseverant on hyping Massey when it's the least relevant.

Massey is virtually meaningless this early in the season. Give it at least 8-10 games to gather current season data please.

I am not hyping them. Indeed, just the opposite. I doubt its relevance at any time. Oregon is a primary reason.

At one point last season, Oregon was ranked above Arizona and UCLA each of whom it had lost to twice PLUS those two teams had better records against the same opponents. AND the same appears true this year. Something is wrong with the algorithm

 

[Sifaka]

I wonder how many Massey Ratings fans and detractors have bothered to read the FAQ on the Massey website. While it focuses on his football ratings, it can probably be used as a good set of generalities for all of his ratings.

Prof.Massey states openly that early season ratings are based on prior years' data, and thus are not particularly accurate. Ditto the predictive value of the matchup tool. As a season progresses, the accuracy of the computer programs improves.

Of course it is easy to state a firmly held, if uninformed, opinion.

Massey Ratings FAQ

 

[visitingcock]

I wonder how many Massey Ratings fans and detractors have bothered to read the FAQ on the Massey website. While it focuses on his football ratings, it can probably be used as a good set of generalities for all of his ratings.

Prof.Massey states openly that early season ratings are based on prior years' data, and thus are not particularly accurate. Ditto the predictive value of the matchup tool. As a season progresses, the accuracy of the computer programs improves.

Of course it is easy to state a firmly held, if uninformed, opinion.

Massey Ratings FAQ

And yet, how did Massey rank UConn ahead of Arizona at the beginning of this year IF it is based on prior years' data? By ignoring last year and giving more weight to 2019-20?

And how did they jump ahead of SCar & UConn? Based upon SCar and UConn finishing ahead of them last year + Oregon finishing one spot above SCar the previous year (by their ranking) and 4 spots above UConn + giving more weight to Oregon thrashing 163rd ranked Idaho State than SCar beating a top 15 (by their rating) NC State? (And "yes" Oregon did admittedly thrash them).

 

[triaddukefan]

All I know is Duke beat a tough Winthrop squad and we have moved up 14 spots to #28. Might be time to break out the hard cider.

 

[Plebe]

All I know is Duke beat a tough Winthrop squad and we have moved up 14 spots to #28. Might be time to break out the hard cider.

By "tough" I assume you mean "tough to watch".

 

[oldhusky]

And yet, how did Massey rank UConn ahead of Arizona at the beginning of this year IF it is based on prior years' data? By ignoring last year and giving more weight to 2019-20?

And how did they jump ahead of SCar & UConn? Based upon SCar and UConn finishing ahead of them last year + Oregon finishing one spot above SCar the previous year (by their ranking) and 4 spots above UConn + giving more weight to Oregon thrashing 163rd ranked Idaho State than SCar beating a top 15 (by their rating) NC State? (And "yes" Oregon did admittedly thrash them).

In terms of Uconn being ranked ahead of Arizona based on last year it actually makes sense. I believe the gap between Uconn's rating and Arizona's prior to their meeting last year was pretty big. After the game it would have logically been much closer, but that doesn't mean you leapfrog over every team you beat.

The rating involves all the games played, and on that basis Uconn may have had a higher rating than Arizona even after the loss because they beat other people more impressively than Arizona did against their competition. As others have mentioned this kind of a system particularly at the beginning of the season takes a number of games to sort things out and be reasonably accurate.

 

[D J B]

All I know is Duke beat a tough Winthrop squad and we have moved up 14 spots to #28. Might be time to break out the hard cider.

At least you'll be good and righteously snockered by the time the real numbers come out.

 

[ConwayGmck]

I wonder how many Massey Ratings fans and detractors have bothered to read the FAQ on the Massey website. While it focuses on his football ratings, it can probably be used as a good set of generalities for all of his ratings.

Prof.Massey states openly that early season ratings are based on prior years' data, and thus are not particularly accurate. Ditto the predictive value of the matchup tool. As a season progresses, the accuracy of the computer programs improves.

Of course it is easy to state a firmly held, if uninformed, opinion.

Massey Ratings FAQ

Our uninformed opinions become more informed as the season progresses. You certainly can't be gracious towards the good Professor, and not be equally gracious towards us, now can you??

 

[Sifaka]

Our uninformed opinions become more informed as the season progresses. You certainly can't be gracious towards the good Professor, and not be equally gracious towards us, now can you??

The uniformed opinions I mentioned were not those about team ratings or rankings. I genuinely respect any fan's views on such matters, with the possible exception of some churlish LV fans.

Rather, I have difficulties with those who lambast a computer program for being bad at something it does not purport to do,
and for failing to read easily available information about the program prior to—or after—skewering it.

My own notions of team rankings are worth just what you pay for them…maybe even less. Yours are likely better.

 

[visitingcock]

The uniformed opinions I mentioned were not those about team ratings or rankings. I genuinely respect any fan's views on such matters, with the possible exception of some churlish LV fans.

Rather, I have difficulties with those who lambast a computer program for being bad at something it does not purport to do,
and for failing to read easily available information about the program prior to—or after—skewering it.

My own notions of team rankings are worth just what you pay for them…maybe even less. Yours are likely better.

I have read the "easily available information about the program." Perhaps I am "skewering" it: Basically, it is the best of the computer programs I have seen though NET ain't 'orful. Nevertheless, it (like human polls) has what seem (to the human mind) to be flaws. Those flaws have led to an overrating of Oregon throughout last year and the beginning of this year. ( I do not assert that Oregon was overranked by Massey at the end of 2019-2020 and admit that my opinion on that subject would be biased.)

I cannot view (and probably would not understand) the entire algorithm. However, I noted throughout last season that Oregon lost every game to the big three of the PAC but Massey consistently ranked them above two of those three - well beyond the ten game "warmup" rankings. They had no excellent wins but 6 "excellent losses." Head to head was not the only problem. Those other two PAC teams also had better conference records in a year when the conference was basically closed to outside competition until the tournament.

One conclusion that I can reach is that the weight of Oregon's 2019-2020 Massey ranking made them highly ranked at the beginning of 2020-2021 and continued to prejudice their ranking (upward) as the season progressed . The fact that the 2019-2020 season is still a consideration in the algorithm TWO years later is troubling - to me. Also, there seems to me to be an overemphasis on MOV - though there are arguments for and against the weight to be given to MOV over hapless opponents. I won't digress.


Oregon may whip our butts in the tournament and I will have to eat some crow THEN. But, at this point, I can only conclude that the algorithm could use some adjustments.

I am fully aware that the swings of yesterday will adjust considerably on Saturday morning. For example, UConn will probably re-jump Oregon with a win.

 

[VaConnFan11]

I think Massey overweights schedule. A lot of times, it looks like a loss to a good team is better than a win over an average team.

This is an example in the Virginia High School football rankings. This team is 2-6 and ranked 10th in the state. They have played a monster schedule, but still.

Life Chr Virginia

2-6 0.250

+2

10 4.81

9 41.08

17 43.26

5 22.45

1.79

1 49.02

1 49.02

0.00

0.0

 

[Glenn]

Having written code similar to Massey's, the aspect of the ratings that seem to get underappreciated by the BY are that it heavily weights point spread. So a win by two , means that you are just a little better than the other team. A win by 20 means that you are a lot better. In his FAQs he say that his algorithm down-weights blowouts--but it does no mean that margin is not important. That said, consistently winning also leads to a high rating. For example, to beat a bunch of teams that are about #25 implies that you are much better than #25, because if you were just a little better, then you would lose frequently. So teams that rack up a lot of wins against mediocre opposition will get ranked highly.

Secondly, even though he uses last year's results, they are used in a "odd way." The rankings make lots of sense, but the team's "power" seems to be pushed towards the median rating. If you look at UConn, for example, it predicts over 5 losses, even though we are the 4th best team. If it really took last year's ratings fully on board it would not be predicting so many losses. That is why first-game blowouts of bad teams have moved good teams way up in the standings. Texas jumped 10 slots to #7 by crushing New Orleans--ranked #352 (who dropped 12 slots) by 131-36. It took a lot of signal from that win as all teams were seen as "kinda" average--even though their rankings were far apart. Again, note how powerful margin of victory is.

 

[visitingcock]

Having written code similar to Massey's, the aspect of the ratings that seem to get underappreciated by the BY are that it heavily weights point spread. So a win by two , means that you are just a little better than the other team. A win by 20 means that you are a lot better. In his FAQs he say that his algorithm down-weights blowouts--but it does no mean that margin is not important. That said, consistently winning also leads to a high rating. For example, to beat a bunch of teams that are about #25 implies that you are much better than #25, because if you were just a little better, then you would lose frequently. So teams that rack up a lot of wins against mediocre opposition will get ranked highly.

Secondly, even though he uses last year's results, they are used in a "odd way." The rankings make lots of sense, but the team's "power" seems to be pushed towards the median rating. If you look at UConn, for example, it predicts over 5 losses, even though we are the 4th best team. If it really took last year's ratings fully on board it would not be predicting so many losses. That is why first-game blowouts of bad teams have moved good teams way up in the standings. Texas jumped 10 slots to #7 by crushing New Orleans--ranked #352 (who dropped 12 slots) by 131-36. It took a lot of signal from that win as all teams were seen as "kinda" average--even though their rankings were far apart. Again, note how powerful margin of victory is.

Thanks.

I don’t like MOV which can lead to RUTS but thanks for your input. Some coaches RUTS when they have poor SOS.

Now, for some alphabet soup.

 

[Dokey]

Well, Massey ranks the other team I root for—Colorado—at No. 16, presumably based on how easily they beat Oklahoma State on the road, and as much as I think Colorado is underrated and quite possibly one of the teams this year who could surprise, I wouldn’t go that far—though that soul crushing defense (11 forced turnovers in first half) blew me—and Oklahoma State—away. But if the offense catches up with the defense...

 

[visitingcock]

Texas now # 2.

UConn falls four spots with win over Arkansas.

 

[FanFromMaine]

You have to love Massey at this time of the season. Teams jump all over the place. And isn’t that what parity is all about???

 

[visitingcock]

You have to love Massey at this time of the season. Teams jump all over the place. And isn’t that what parity is all about???

I will admit it’s fun to see teams rise and fall 65 places based in a game or two. Yes, yes those jumps and falls will be less glaring when more games are played.

 

[Charliebball]

as is often said, the only poll that really matters is the one that comes out directly after the national championship game has been played......

 

[visitingcock]

as is often said, the only poll that really matters is the one that comes out directly after the national championship game has been played......

Even the polls don't really matter. The trophy matters and that means the number one ranking (though theoretcially it might not). Afre that number one ranking, the rest doesn't matter that much.

And as others would point out, Massey isn't a "poll."

 

[Artifact]

Oregon Jumps to Number 2. UConn falls to 4. NC State falls 10 spots. Texas rises 10 spots.. Massey Ratings - College Basketball Women's : NCAA D1 Ratings
Ratings expressed after only one game played? Never heard of this. ' Massey is a Mess'

Stanford Pac 12	0-0 0.000		1 9.32	1 69.34	4 100.02	3 33.99	2.90	2 69.34	1 55.42	22.94	6.06
Oregon Pac 12	1-0 1.000	+4	2 9.15	2 67.78	10 97.77	1 34.67	2.99	171 36.82	9 52.88	21.01	5.99
South Carolina Southeastern	1-0 1.000	+2	3 9.09	4 66.02	12 97.71	5 32.97	2.90	14 61.48	6 53.12	20.40	5.60
Connecticut Big East	0-0 0.000	-2	4 9.02	3 66.69	5 99.30	6 32.05	2.90	5 66.69	25 51.17	23.70	5.30
Maryland Big 10	1-0 1.000	-1	5 8.88	5 65.42	1 104.87	47 25.22	2.87	252 27.02	16 52.09	21.36	7.64