Page 5 of 7 FirstFirst ... 34567 LastLast
Results 61 to 75 of 93
  1. #61

    Default

    Quote Originally Posted by howeda7 View Post
    Not sure it a matters, but they're #32 in the AP poll this week. If they would close 4-1 or something, they will probably be ranked in the top 25. Not many top 25 teams get left out, even if their RPI is still poor.
    Only one person voted for the Gophers, and it’s hard to say how much they pay attention. Saying they rank #32 to the AP (in this case, one person), isn’t very meaningful.


  2. #62

    Default

    Quote Originally Posted by hungan1 View Post
    Just keep winning. It will speak volumes.
    Bingo. We can do all the math we want. Just win.

    Avenge Purdue. Demolish Penn State
    Shock Maryland. Sweep Rutgers.
    Avenge Michigan State.
    This team has tools to beat ANY big ten team.
    The key is using the press to make teams uncomfortable, and keeping the extra shooter on the court to open the paint against big ten defense

  3. #63

    Default

    Quote Originally Posted by Ignatius L Hoops View Post
    ... found a few power five conference teams that made the tournament with RPI's in the 60's. I'm assuming that those teams were from the highest ranked conference. Right now the B1G is the 4th ranked conference. ...
    I didnít know they ranked conferences. Thatís intriguing. Do you have a URL where they give conference rankings?

    I googled, but Google is so thick-headed that as soon as I use the terms basketball and tank, it just blithely assumes I want team rankings, and thereís no way to convince it otherwise.

  4. #64

    Default

    I tried a different search engine but only found rankings for Men's conference
    Last edited by EaganGopher22; 02-13-2019 at 08:04 AM.

  5. #65

    Default

    Quote Originally Posted by CutDownTheNet View Post
    I didn’t know they ranked conferences. That’s intriguing. Do you have a URL where they give conference rankings?

    I googled, but Google is so thick-headed that as soon as I use the terms basketball and tank, it just blithely assumes I want team rankings, and there’s no way to convince it otherwise.
    Here's the NCAA main RPI page. Just fill in Division, Sport and Gender a list will appear below.

    https://extra.ncaa.org/solutions/rpi...ages/Home.aspx

    The conference RPI is listed on the various sheets. Here's the nitty-gritty sheet from yesterday for instance:

    https://extra.ncaa.org/solutions/rpi...%202-12-19.pdf

  6. #66

    Default

    Quote Originally Posted by Ignatius L Hoops View Post
    Here's the NCAA main RPI page. Just fill in Division, Sport and Gender a list will appear below.

    https://extra.ncaa.org/solutions/rpi...ages/Home.aspx

    The conference RPI is listed on the various sheets. Here's the nitty-gritty sheet from yesterday for instance:

    https://extra.ncaa.org/solutions/rpi...%202-12-19.pdf
    Thanks!

  7. #67

    Default

    For those scoring at home, Gopher opponent Arkansas-Pine Bluff has the fifth-worst RPI in the nation. (And, since it's wins and losses that actually affect the Gopher RPI, not opponent RPIs, it's worth noting that Pine Bluff has a grand total of one win. Schedule-wise, they're the gift that keeps on taking.)

  8. #68

    Default

    https://extra.ncaa.org/solutions/rpi...%202-17-19.pdf

    B1G RPI through games of 2-17-19 with SOS and conference record (we remain above the really terrible B1G teams and below the remainder):

    9 Iowa 11 (12-3)
    17 Maryland 100 (12-3)
    27 Rutgers 30 (9-4)
    29 Michigan State 46 (7-7)
    48 Indiana 51 (6-8)
    49 Michigan 78 (9-6)
    60 Purdue 31 (8-7)
    85 Nebraska 20 (7-8)
    93 Ohio State 39 (8-7)
    96 Northwestern 92 (8-7)
    102 Minnesota 203 (8-7)
    136 Penn State 25 (3-11)
    184 Wisconsin 152 (3-11)
    191 Illinois 91 (2-13)

    Non-conference Gopher opponents:

    12 Syracuse
    117 Boston College
    192 Xavier
    227 Cornell
    236 Rhode Island
    259 San Diego
    275 Air Force
    308 Incarnate Word
    317 New Hampshire
    329 Coppin State
    346 Arkansas Pine Bluff

  9. #69

    Default

    Gophers are #39 in Massey. Not sure what criteria is different between the two but that is a large differential when comparing to RPI.

  10. #70
    Join Date
    Nov 2008
    Location
    Twin Cities
    Posts
    3,030

    Default

    Quote Originally Posted by Waffle View Post
    Gophers are #39 in Massey. Not sure what criteria is different between the two but that is a large differential when comparing to RPI.
    Must be strength of schedule

  11. #71

    Default

    Quote Originally Posted by Donovan View Post
    Must be strength of schedule
    Massey Ratings

    https://www.masseyratings.com/cbw/ncaa-d1/ratings

    through Feb 17, 2019 games
    (Includes leader Baylor as a point of reference)

    Team Rec Rating Off Def SoS
    Baylor 23-1 2.74 (1) 105.21 (7) 36.51 (1) 55.37 (5)

    Maryland 23-3 2.32 (10) 100.99 (15) 29.64 (31) 49.77 (42)
    Iowa 21-5 2.26 (12) 105.04 (8) 24.00 (86) 53.17 (14)
    Rutgers 18-7 2.04 (28) 85.71 (139) 35.23 (3) 50.52 (30)
    Michigan 18-9 1.98 (34) 93.94 (47) 26.60 (54) 49.72 (43)
    Michigan State 17-8 1.97 (36) 99.36 (21) 21.89 (121) 50.24 (35)
    Minnesota 19-7 1.95 (39) 92.12 (64) 28.35 (39) 46.41 (72)
    Purdue 17-11 1.89 (50) 88.11 (106) 29.87 (29) 50.98 (26)
    Indiana 17-10 1.84 (57) 93.28 (52) 24.77 (74) 49.99 (38)
    Northwestern 15-11 1.84 (59) 89.56 (84) 29.67 (30) 49.67 (44)
    Ohio State 12-12 1.82 (61) 87.77 (109) 27.63 (47) 52.52 (18)
    Nebraska 12-14 1.81 (62) 96.36 (34) 22.70 (109) 52.70 (15)
    Wisconsin 12-14 1.54 (117) 86.34 (127) 23.33 (103) 45.66 (79)
    Penn State 10-15 1.52 (122) 92.51 (57) 17.03 (209) 48.53 (53)
    Illinois 10-16 1.43 (138) 89.37 (88) 17.66 (195) 45.98 (75)

    Non-conference Gopher opponents

    Team Rec Rating Off Def SoS
    Syracuse 19-6 2.16 (18) 100.44 (17) 25.66 (65) 52.52 (17)
    Boston College 14-12 1.64 (93) 97.88 (29) 14.65 (249) 46.12 (74)
    Xavier 10-14 1.44 (134) 83.16 (177) 23.55 (97) 44.94 (84)
    Cornell 9-10 1.19 (200) 75.61 (285) 23.62 (95) 34.85 (211)
    San Diego 8-18 1.03 (243) 78.68 (242) 14.21 (252) 39.57 (123)
    Air Force 8-16 1.01 (248) 78.38 (246) 16.26 (219) 37.04 (180)
    Rhode Island 7-17 1.01 (250) 80.31 (221) 12.93 (274) 38.56 (150)
    New Hampshire 6-20 0.65 (319) 69.53 (338) 13.96 (256) 29.77 (287)
    Incarnate Word 3-20 0.57 (324) 73.47 (310) 8.73 (319) 33.48 (231)
    Arkansas Pine Bluff 5-14 0.34 (346) 63.90 (350) 11.19 (299) 21.30 (349)
    Coppin State 2-22 0.28 (349) 72.33 (316) 4.57 (343) 29.06 (298)

    The numbers in parens are rankings for the raw ratings to their left. The Rating (Ranking) and SoS stats loosely correspond to RPI and Strength of Schedule (and their rankings) only on a different scale.

    I’m not sure, but the Off(ense) and Def(ense) raw ratings seem to loosely correspond to (respectively) number of average points a team would score against a zero-rated defense, and decrement one should subtract from other team’s offense.

    The key point about Massey ratings is that it is most likely the most mathematically and statistically sophisticated set of metrics, and due to that it is probably the least (improperly) biased metrics. It can be read as being closest to the best estimate of team basketball quality.

    Compared to Massey as a good source of truth, the bad to worse metrics stack up as follows:

    RPI is not a direct measure of team basketball quality. Rather, it is mostly a measure of how good a job a team’s administrative staff did in lining up non-conference teams to play that have pretty high rating yet are still beatable. Oh, and just for grins they throw in 1/4 times a measure of the won-loss record, adjusted for home advantage. Last year’s Gopher staff did a horrible job at playing this game, so we get assigned an RPI of #102 at the moment. Minnesota loses the administrative RPI game. Other teams win the RPI game by lining up top-ten teams to play, and they gain RPI creds by just playing and losing to them. In other words, RPI is extremely tilted toward putting an over-emphasis on strength of schedule. Also, RPI only counts who won and where, discounting completely margin of victory.

    On the other extreme, the ugly new men’s NET rating is equally bad (or possibly worse). It is mostly based on margin of victory (not just who won), and places too little emphasis on strength of schedule. So you may see weak teams making it to the playoffs just because they killed other weak teams in margin of victory. And teams will try to game NET by leaving starters in just to increase margin of victory. Also, one part of NET violates a fundamental principle of statistics that is too deep to get into here.

    At least with RPI, if you're reasonable in setting up your NC schedule, you can keep RPI from killing you and putting you on the NCAA-tournament scrap heap. Stollings left us a horrible NC schedule in her will before leaving, and so we pay the price for not setting up a reasonable schedule. With NET on the other hand, a team is essentially encouraged to game the NET system. So rather than the default of (just do a good job scheduling your NC schedule) that exists with RPI, instead with NET the schools have an insidious motivation to schedule weak schools in their NC schedule - just so that they can get lots of games in which they run-up the margin of victory. So once schools realize this "feature" of NET for the men's game, we may see schools rushing to schedule Coppin State and Arkansas Pine Bluff into their NC schedule. In fact, quite ironically, actually, if the NCAA should happen to make the huge mistake of switching the womens' game from RPI to NET starting next year, we'd actually have to send an emergency telegram to the Gopher Womens' Basketball coaching staff to "hold the horses on dismissing Coppin State and Arkansas Pine Bluff from our NC schedule, since now that NCAA has flipped the scheduling game that we must play, we (like the men) now actually want to load up with cupcake teams! NCAA is so stupid.

    Other metrics are available. KenPom is well known. But I’ve read that it is 98% correlated to NET, and thus inherits the NET evils.

    So Minnesota Lady Gophers get a Massey Rating that ranks #39 on the Massey scale. That’s a pretty fair ranking. No doubt it’s hurt just a little bit by weak NC schedule - but not hurt much, since Massey has built in math features that recognize much of our weak NC schedule as throw-away games, and thus minimize the negative impact thereof, as opposed to RPI that is dominated by our weak schedule, and in fact is mostly a measure of Schedule weakness or strength. In fact, to that point, Massey's SoS ranking for Minnesota is #72. That's weak, but a far cry from the #203 SoS ranking given to us by RPI. That's because Massey minimizes the impact of that weak schedule on its Massey rating system, whereas RPI maximizes the impact of that weak schedule in its RPI.

    The amount of SoS impact that remains (even in Massey, which downplays it at #72) from our cupcake NC schedule - well, that's pretty much what puts us in 6th Massey place among the Big Ten, right behind Rutgers, Michigan and Michigan State. If we had had control over our NC schedule this year, and if the new coaching staff had utilized that control to do a better NC scheduling job, we could have easily scheduled some #150 to #200 teams in place of the #250 to #350 teams, and we still could have beat those substitute teams, so we'd have the same record against a vastly improved (in RPI's mind) SoS (and a moderately improved Sos in Massey's mind). That would have almost certainly put us in 3rd place among Big Ten teams on the Massey scale.

    From the reasoning in the last paragraph, and given that Massey is a more accurate metric of "how good a basketball team that a given team is," we can at least interpret that with a bit of pride in that it states that the Lady Gophers are approximately the 3rd-best team in the Big Ten. And on any given game day, we're also capable of beating Maryland and Iowa. So we only need to keep up the trend of excellence, in an effort to overcome the clerical error that is our current RPI ranking.

    So in a word, yes, the difference is that Massey (properly) minimizes the impact of poor SoS, whereas RPI gloats on poor SoS.
    Last edited by CutDownTheNet; 02-19-2019 at 08:23 PM.

  12. #72

    Default

    In slightly good RPI news, Arkansas Pine Bluff halted a seven-game losing streak last night.

  13. #73

    Default

    Quote Originally Posted by thatjanelpick View Post
    In slightly good RPI news, Arkansas Pine Bluff halted a seven-game losing streak last night.
    Sweet! Nice job, APB.

    Holy RPI !!! That bumps the Gophers RPI from #102 up to #101, according to RealtimeRPI. And our SoS jumps from #202 all the way up to #149. I’m guessing some of our other opponents besides APB might also have won a game.

    And Gamer’s prediction for our end-of-season RPI is now #89 with a final estimated SoS of #111. This is based on Gamer thinking we’ll lose our two road games at Maryland and Rutgers, and win our home game with MSU.

    Of course, that means we can do better than a final #89 RPI if we can steal one of those road games (or preferably both) while at the same time holding serve against Michigan State.

    Go Gophers!

  14. #74

    Default

    [Warning: This got to be a longer rant than any normal rant, even by my standards. So caveat reader. But my line of thought led me to what I actually think is a good solution to the problem of bad RPI and bad NET rating systems, and bad selection algorithms used by the Selection Committee.]

    An interesting couple of side notes regarding Massey basketball ratings, especially as compared to RPI and NET ratings.

    On the one hand, Massey ratings are fair and mathematically/statistically sophisticated and thus about as close as any model can get to a source of truth re the question, “How good are each of this set of teams at playing basketball, relatively speaking?”

    Compared to a baseline Massey “quite accurate source of truth,” the RPI metric is highly skewed in the direction of putting way too much emphasis on a team needing to play against a set of comparable teams, and severely punishing a “Minnesota Nice” team that generously offers to play a set of teams way below its caliber. Is Minnesota Nice such a crime that it deserves the punishment of effectively banning the Lady Gophers from the NCAA tournament? I think not. Massey treats that Minnesota Nice generosity with fairness, whereas RPI treats it as a capitol offense.

    The NET metric tries to remedy the symptom of RPI hyper-emphasizing having a quality Strength of Schedule (and not playing too many cupcake teams). It does achieve that goal, but at what a cost! If you consider the analogy of RPI being a sinking boat capsizing by taking on water on the port side, then NET is a sinking boat capsizing by taking on water on the starboard side. By switching the Men to NET we have swapped out one sinking ship for another.

    Furthermore, the way NET works is now giving bonus credits for playing cupcake teams. Because when you play cupcake teams you increase your average margin of victory, and that’s a key factor in improving your NET rating. If we thought that playing cupcake teams was such a bad idea in Womens hoops that we were OK with giving the Minnesota Nice Lady Gophers the NCAA death penalty for doing so, then why did we switch over the Mens rating system to a NET rating that absolutely encourages teams to schedule cupcake teams into their NC schedule so as to game the NET rating system to their advantage. All of a sudden we’re going to have 50 states worth of Mens teams all rushing to be Minnesota Nice (and probably offering cash cupcake bribes to any cupcake team that is willing to play them. I can see it now, there will probably be a cupcake black market on the Dark Web.)

    By the way, the NET rating system came from assembling a committee of Mens basketball coaches and letting them make their political arguments, and then come to a consensus recommendation. All of these committee members were statistical idiots. Not to mention mathematical illiterates. They had to have been, given that what they came up with was just as bad as RPI, only failing in the other direction. To add insult to injury, to get the advertising benefits, the otherwise technically sound Google signed on to help the committee to commit an egregious Data-Science error. NET is a comedy of errors that would be difficult for a fiction writer to concoct.

    But consider the following proposal for a completely revamped rating system and Committee selection criteria, which would completely solve the problem.

    Instead of a boat swamping on the left (for the Women) and a boat swamping on the right (for the Men), there is a simple solution to the current Fiasco of the NCAA using two statistically and mathematically bankrupt metrics designed by committees of statistically illiterate coaches, to try to sort out who makes the playoffs at the end of each season.

    The better idea goes like this (but of course it’s too sensible for the NCAA’s likings):

    You start by using a mathematically and statistically sound and fair metric as your primary judge of basketball quality - with Massey ratings being one good exemplar of such. You also bump the field to 96 teams from 64, each of the 96 teams being ranked, mostly (see below) on the basis of the new (Massey-like) basketball-quality rating metric, but some discretion allowed for the Committee. The 65-96 ranked teams will compete in an extra preliminary round against the 33-64 ranked teams, thus determining the 33-64 placed competitors who will compete in the round-of-64 as usual.

    Admission to the initial 96 candidates is as follows: First, each conference’s winner is added to the list of 96 (however that winner is determined, such as conference tourney for Big Ten). Let’s say that K slots remain to be filled. You choose each of those K teams by going right down the (Massey-like) ranking list starting with first place. If a team is already on the list (by virtue of winning the conference) then skip it cuz it’s already on the list, otherwise add it to the list. Repeat until you select a total of 96 teams. When you get to selecting the last 16 teams, the Committee gets some discretion, and may deviate from list order due to other resume factors.

    This way, not only are there some extra slots available for second-place teams in also-ran conferences (in case that conference had a great year), but there are also additional slots available for Power-5 teams that otherwise would have been cut in order to make room to give each conference at least one slot on the NCAA tournament. Let these marginal teams play their way into the tournament.

    This better system could be used for both Men and Women NCAAs. After all, both RPI and NET are messed up (in different ways), so both RPI and NET need to be replaced. It’s absolutely not a matter of determining which of RPI or NET is “best” in some sense, and choosing to use that one for both. Both of them are worst in their own sense.

    If we had used the above sensible system this year to let supposedly marginal teams play their way into the tournament, then here’s how it would look for the Big Ten (using the current almost-end of-season Massey ratings shown in parentheses) ...

    Maryland (#10)
    Iowa (#12)
    Rutgers (#28)

    Michigan (#34)
    Michigan State (#36)
    Minnesota (#39)
    Purdue (#50)
    Indiana (#57)
    Northwestern (#59)
    Ohio State (#61)
    Nebraska (#62)

    The first group of three (Maryland, Iowa, Rutgers) would get a bye for the play-in round.

    The second group of eight (Michigan, MSU, Minnesota, Purdue, Indiana, NW, OSU, Nebraska) would all get an opportunity to play their way into the round-of-64 via the new play-your-way-in round - with each receiving matchups based on their relative rankings.

    You may have noticed that under this better system fully 11 out of the 14 Big-Ten teams get a chance to compete in the NCAA playoffs. This is indicative of how strong the Big Ten is, as well as how tightly grouped it is in the middle of the pack this year, thanks to parity in the middle. The three Big Ten teams not getting a chance under the proposed new system are ...

    Wisconsin (#117)
    Penn State (#122)
    Illinois (#138)

    I think we can agree that those three are not bubble teams this year.

    But after the first-three definitely-in teams, every one of the 8 middle-of-the-pack teams is bubble-ish in at least some sense. In the current horrible RPI-based selection system, the Committee is left to chose from among these 8 teams in what amounts to a very arbitrary manner. They might say something like “Let’s cross the Minnesota Nice Lady Gophers off the list as punishment for playing too many cupcake teams” - even though everybody knows that they are just as good as the Michigans and better than Purdue/Indiana/NW.

    Continuing their conversation, the Committee might say something like “Whew, that was easy, now what arbitrary criterion are we going to use to cut the remaining 7 Big-Ten pack teams down to, say 3, or maybe 4 if we’re generous?”

    [And meanwhile, across the hall in the Mens Committee meeting, they’re using the equally ridiculous NET rating to decide which teams to let In the playoffs, based on the fact that they played a sufficiently high number of cupcake teams to artificially jack up their NET scores. I’m not making this up people, this is exactly what will be happening in a few weeks. One can’t help but note the irony in this.]

    Playing your way into the round-of-64 is just a better idea in so many ways.

    The following non-conference Gopher opponents would also find themselves with an NCAA opportunity under the proposed new system ...

    Syracuse (#18) with a bye for the play-in round
    Boston College (#93)

    So even Boston College gets a hard-fought opportunity to make it into the NCAA tourney.

    This idea is partly inspired by how well it has worked for the WNBA to instate one-and-done games as WNBA tournament qualifiers.
    Last edited by CutDownTheNet; 02-20-2019 at 12:25 AM.

  15. #75

    Default

    It's not just the non-conference schedule. The conference losses (and schedule) are also a major drag on Minnesota's RPI. Minnesota's only top 50 RPI conference win is Rutgers. We've played and lost to Iowa, Michigan and Michigan State (Indiana's RPI rose above 50 with the loss to Rutgers.

    Yes, the next three games could improve the conference ranking. But, currently, four of our 15 games were against Penn State (3-11), Illinois (a 2-13 loss) and Wisconsin (twice at 3-11). Other than Rutgers the only conference teams with winning record that we've defeated are 8-7 (Northwestern and Purdue). In a ranking system that only cares about wins and losses that doesn't move the needle.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •