Bubble Breakdown: Where Marquette Really Stands
Written by: noreply@blogger.com (brewcity77)
CS Team prologue: Please welcome Brewcity77, the newest member of the CS writing team!
---
On my way out of the Cincinnati game, I heard a Marquette fan behind me saying "Well, now they need to beat Seton Hall, and probably get two in the Big East tournament to get in." Unable to help myself, I interjected "I really don't think so. One win would certainly be enough, though they may have a good enough resume even if they lose out." A brief discussion followed as we exited, and I think I reassured him that Marquette's situation isn't as dire as many fans and online gurus may indicate. When you look at Lunardi's Bracketology, Marquette still hasn't fallen to his Last Four In. Before the Seton Hall loss, the general consensus was that Buzz's team was in as a 10-seed. Even if that loss drops them a seed line, they should still be in.
However, there are a number of worrying criteria that do cast a shadow of doubt on Marquette's resume. RPI, total losses, all of these factors swirl around the collective heads of the Marquette team like a grim specter of doom. I've been of the opinion that these factors aren't as dire as many say for quite awhile, and to prove (or disprove) my own feelings, I did a breakdown of Marquette and 29 other teams that are either considered securely in, on the bubble, or on the outside looking in. The criteria I used were current RPI, projected Strength of Schedule (hereafter SOS), out-of-conference (hereafter OOC) RPI, out-of-conference SOS, wins against RPI top 50 opponents (quality wins), losses against RPI 101+ opponents (bad losses), the plus/minus rating comparing quality wins and bad losses (all taken from RPIForecast.com (http://www.rpiforecast.com/live-rpi.html)), total current losses, and ratings from both kenpom.com (http://kenpom.com/) and Sagarin (http://www.usatoday.com/sports/sagarin/bkt1011.htm). Of the 30 teams I compared, 18 should get bids, meaning that for Marquette to be "worthy" in a category, they need to be in the top 18.
The teams I am comparing Marquette to include all projected at large teams from the latest S-Curve (http://www.bloggingthebracket.com/2011/3/4/2029373/bracketology-ncaa-tournament-march-madness) and includes the following: Utah State, Old Dominion, Tennessee, UCLA, Michigan State, Georgia, Illinois, Virginia Tech, Gonzaga, St. Mary's, Butler, Clemson, Boston College, Michigan, Richmond, Missouri State, UAB, Alabama, Colorado State, Colorado, Oklahoma State, Washington State, Penn State, Virginia Commonwealth, and Baylor. I also included USC, Memphis, Nebraska, and UTEP because they were recently added to Lunardi's Last 8 Out (http://www.muscoop.com/index.php?topic=25166.msg280465#msg280465). While some of these teams, notably Utah State, Gonzaga, St. Mary's, Butler, and Missouri State, are in a position to qualify via an automatic bid, I included them because they will likely be compared to Marquette if they fail.
With 30 teams, I have also broken down the categories into three color classifications. Red means Marquette is in the bottom ten of the category and that it is bad mark on their resume. Orange means we are in the middle ten, which will likely yield 5-8 bids, but while not a negative also isn't an overwhelming positive. Green means Marquette is in the top ten of the category and that it is a positive on their resume. The numbers I am using have been taken on the morning of March 6. Now, finally, to the numbers...
RPI (24/30)
Marquette's RPI of 68 ranks 24th of the 30 teams and is one of their most troubling numbers. A slight positive is that there are 4 teams ahead of them that are also in the 61-67 range, which is comparable to Marquette's rating and would put them in the top 20, just outside the bid-worthy range. If they managed to win two games in the Big East tournament, that would move them up to an approximate 55 RPI, which would put them at 17th in RPI, right around where they need to feel somewhat secure. If they win one game, they will likely stay around where they are at, or possibly climb one or two spots. If they lose to Providence, they will fall to around 80, which would leave them at 28th of the 30 teams and in severe danger in this category.
SOS (6/30)
Marquette's SOS of 31 ranks 6th of the 30 teams and is a definite positive. Despite what was considered a weak OOC schedule, the grueling nature of the Big East was such that Marquette has gone through one of the tougher roads in the country. And no matter what happens in the Big East tournament, this isn't likely to change much. If we lose to Providence, we might drop a couple places, but certainly not enough to drop us out of the top ten of the 30 teams in this study.
OOC RPI (28/30)
This is even more damaging to Marquette than their actual RPI. Marquette ranks 28th of the 30 teams with a 122 OOC RPI. Playing a weak schedule, coupled with losses in their four biggest games, was very damaging. The only positive is that Bucknell and UW-Milwaukee both played their way into the top 100 RPI, giving them a pair of decent non-conference wins, even if they won't count as true "quality wins".
OOC SOS (20/30)
This one surprised me considering how much has been made of our weak non-conference schedule, but Marquette checked in right at 20th of the 30 teams. A little bit outside the range of where we'd like them to be, but still at least respectable. While this won't help them, it also isn't nearly as bad as we might have expected.
Quality Wins (T4/30)
Only 3 teams did better than Marquette in this category. Tennessee has a very impressive 7 quality wins, while both USC and Colorado are tied for second with 5. Marquette is tied with Memphis and Butler at 4 apiece, but it may be important to note some factors for those two teams. Three of Butler's quality wins come against the same opponent, Cleveland State. The other comes from a Florida State team that sits at 47 in the RPI, a position that could see them fall out of the quality win territory. For Memphis, two of their quality wins come from Southern Miss, a team that is teetering right on the 50 mark. Any falloff by those Golden Eagles could easily see Memphis' four quality wins reduced to two. Either way, Marquette is in a great position in this regard.
Bad Losses (T1/30)
Marquette and Georgia are the only two teams without a single bad loss in the bunch. While most viewed Seton Hall as a bad loss, the Pirates' RPI was already at 97 after they beat St. John's, and climbed to 91 with the win over Marquette. This is a great mark for Marquette and one that will certainly be one of the most favorable points that their supporters in the Selection Committee can make.
Plus/Minus (T1/30)
It's easy to disregard this stat as a simple composite of the two previous stats, but only Marquette and Tennessee have a +4 differential. What makes this significant is that when you look at teams like Colorado (+2) and USC (-1), their 5 quality wins seem a lot less striking when compared with the number of times they were tripped up by "bad" opposition. When you look at the relative quality of Marquette's wins coupled with the quality of the opponents that beat them, it shines brightly in our favor.
kenpom.com Rating (5/30)
How much will the computer ratings play in to the Selection? It's hard to say, but we do know that the Committee is provided both the kenpom and Sagarin ratings. They may not be relied on as heavily as the RPI is, but Marquette's tendency to keep their margin of defeat narrow ensures that they will always be ranked much higher in these ratings than they will be in the RPI.
Sagarin Rating (3/30)
Similar to the previous rating, Sagarin also finds great favor with Marquette. With two exceptions, Marquette only loses to very good teams, and with two exceptions, Marquette only loses by close margins. Score another big point for Buzz's boys.
Total Losses (T27/30)
First, the reason I include current total losses rather than including the expected loss that would make these teams bubble teams is to keep everyone equal. Any team on the list that manages to go without a loss will be in regardless, so I'm sticking with current numbers. Clearly Marquette's 13 losses won't be a positive, as it is tied for worst of the 30 teams. And if Marquette does make the tourney as an at-large, they will obviously go in with 14 losses. However, this category is one where we may benefit from the Committee members that use RPI as their primary factor in picking teams. 11 of Marquette's 13 losses come to teams in the top 35 of the RPI. Of the other 29 teams, only one even played against that many top 35 RPI teams, that being Georgia, who played 12. So while this will likely be a negative, Marquette can draw some solace out of the knowledge that this could turn into a positive for those that focus strictly on the RPI.
Aggregate Numbers
Quite simply, that's a lot of data. Some of it favors Marquette, some of it could work against them. So I suppose the bottom line is where they rank amongst these 30 teams that are scrapping for the last 15-18 bids the Selection Committee will hand out. While it's unlikely that all of these factors will be considered equally by the Committee, that's what I did to determine where these teams stood. I assigned each category a point value from 1-30, with the top team getting 30 points per category, ranging down to 1 point for the bottom. In cases of ties, I assigned all the tied teams the same amount of total points. Marquette finished 12th of the 30 teams. I actually expected them to be higher, but huge hits from their RPI and total losses keep them on the bubble, but seemingly on the right side.
(http://2.bp.blogspot.com/-CNs5U9vYZFo/TXQGISrug4I/AAAAAAAAAA4/Wsu9vDwTrW0/s400/MU%2BSheet.jpg)
(http://2.bp.blogspot.com/-CNs5U9vYZFo/TXQGISrug4I/AAAAAAAAAA4/Wsu9vDwTrW0/s1600/MU%2BSheet.jpg)
http://www.crackedsidewalks.com/2011/03/bubble-breakdown-where-marquette-really_06.html
Best post ever? Thanks for putting this together, it was info I was curious about but too lazy to look up on my own. Awesome work.
Does this mean we are in good or bad shape?
Thanks, suffice to say, it took awhile :D
And I'd also say that a few of these teams are already locks. UCLA and Old Dominion are consensus 8 seeds and appear in all 78 of 50webs.com (http://bracketproject.50webs.com/matrix.htm)'s multi-bracket projection from earlier today. Utah State, Illinois, and Tennessee are also all likely locks. They appear in all 78 brackets as consensus 9-seeds and none of them are listed lower than an 11-seed in any of them. That leaves most likely 10-13 bids for the remaining 25 teams.
So what does that mean for Marquette? Well, we appear in 76/78 brackets as a consensus 10-seed. Based on my numbers, of the non-locks, Marquette would qualify as the 7th best team of the remaining 25. In addition, the inclusion of Tennessee and Illinois as locks is a huge boost to Marquette's chances. Assuming neither of them earn automatic bids, they will finish with 13 losses, one more than Marquette is projected to finish with. In addition, both finished their regular conference seasons with .500 records in conferences that don't require nearly as brutal a slate as the Big East offers.
I maintain that a loss to Providence would be damaging, and leave us all sweating next Sunday, but would probably see us still make the field, and personally I don't think we'd be in the play-in game unless everything went right for teams behind us in the pecking order. If we beat Providence, however, we become a lock. Our profile is simply too good, despite our RPI.
Thanks for this post! A lot of work to do!
Great post!! Now we can send all the "there is no chance Marquette makes the field" idiots here.
Plus/Minus is my new favorite stat ;)
Quote from: 6Under20 on March 06, 2011, 06:48:22 PMPlus/Minus is my new favorite stat ;)
That's what I get for years of growing up with NHL video games ;)
We hear so often about quality wins in relation to bad losses, but none of the pundits ever really put it into words. When I looked at the relative comparison of the two, especially with teams like Marquette, who had 4 quality wins and a +4 rating, and USC, who had 5 quality wins but a -1 rating, I think it's an incredibly important stat. And even if the Selection Committee and pundits don't list it as a category, I find it very hard to believe that any reasonable SC member wouldn't look at quality wins and bad losses in relation to each other. I just wanted to put a label on it ;D
FWIW, Statsheet has MU in the second group of four teams as out of the NCAA's.
http://goldeneaglefan.com/marquette-basketball/statseed-update/statsheet-projects-marquette-as-ncaa-tournament-bubble-team-03-06-2011#
Quote from: Dr. Blackheart on March 06, 2011, 06:58:58 PM
FWIW, Statsheet has MU in the second group of four teams as out of the NCAA's.
http://goldeneaglefan.com/marquette-basketball/statseed-update/statsheet-projects-marquette-as-ncaa-tournament-bubble-team-03-06-2011#
If you look at the teams that are around them in "last 4 out" and "next four out" it becomes clear that this service values W-L record and RPI more than anything else. We have a resume that is volumes better than OK State, Marshall, Southern Miss, Cal, Miami (FL) and USC. In fact, it's not even really close.
Marquette will have to play their way into the Tournament because as of now, in my opinion, they're on the outside looking in. Beating PC, which I don't see as an easy task, is crucial and then vs. WVU will be the true test if they're tourney worthy. Anything less and I think the season is curtains.
Quote from: JamilJaeJamailJrJuan on March 06, 2011, 07:06:33 PMIf you look at the teams that are around them in "last 4 out" and "next four out" it becomes clear that this service values W-L record and RPI more than anything else. We have a resume that is volumes better than OK State, Marshall, Southern Miss, Cal, Miami (FL) and USC. In fact, it's not even really close.
Agreed. Looking at it, I just can't see it. Us behind Minnesota and Oklahoma State, which had absolutely awful conference records? Cleveland State in the tournament with no top 50 wins and 20/24 wins outside the top 100? I can't remember many bubble teams getting at-large bids without any quality wins.
I know RPI is important, but it's not the only factor the SC looks at, which is the reason why we have any chance whatsoever.
Quote from: 4everwarriors on March 06, 2011, 07:07:57 PMMarquette will have to play their way into the Tournament because as of now, in my opinion, they're on the outside looking in. Beating PC, which I don't see as an easy task, is crucial and then vs. WVU will be the true test if they're tourney worthy. Anything less and I think the season is curtains.
And that's exactly why I wrote this. To show where we actually stand in terms of measurables compared to other NCAA bubble teams. I tried as much as I could to keep personal opinion out of this article. If you think Marquette is on the outside looking in, please offer concrete evidence as to why this is, not just another shot at the team with no basis in fact.
Quote from: JamilJaeJamailJrJuan on March 06, 2011, 07:06:33 PM
If you look at the teams that are around them in "last 4 out" and "next four out" it becomes clear that this service values W-L record and RPI more than anything else. We have a resume that is volumes better than OK State, Marshall, Southern Miss, Cal, Miami (FL) and USC. In fact, it's not even really close.
Not attesting to this one way or the other but just another statistical view run through their models. Here is the link on how calculated and a snippet I cut and pasted.
http://statsheet.com/blog/statseed-projecting-every-team-for-every-postseason-tournament
HOW WE SEED THE TEAMS
We create a StatSeed for every team in every tournament, and publish those seeds for the NCAA Tournament and NIT teams. The seed is determined by a combination of factors and calculations, taking into account rankings like the AP Top 25 and the RPI, conference standings, conference strength and selection history, regular season winning percentage, projected conference tournament finish, and records against top competition (quality wins and bad losses).
Thus, our final at large bids appear in the 9-13 seed range, while automatic bids from weaker conferences fill out the lowest seeds. We also take into account the fact that some of our projected NCAA BUBBLE teams will wind up in the NIT, some of our NIT BUBBLE teams will wind up in the CBI, and so on.
Quote from: Dr. Blackheart on March 06, 2011, 06:58:58 PM
FWIW, Statsheet has MU in the second group of four teams as out of the NCAA's.
http://goldeneaglefan.com/marquette-basketball/statseed-update/statsheet-projects-marquette-as-ncaa-tournament-bubble-team-03-06-2011#
Well that's not good.
Also, the Bracket Matrix does have us in 76 of 78, but over 80% of them were last updated on 3/4 prior to our crushing at Seton Hall. Tomorrow is when most are updated.
Brewcity, thanks for a great post. I don't even want to think about losing to Providence, but if we do, my concern is that we will be going before the committee with three straight losses, with one at home and then two losses to non-tournament teams.
It seems like committee spokespersons have been somewhat vague about how important the "last 10" record is (particularly in the absence of injuries), but you have to think a three game losing streak against these opponents means more in March than it would in January. If we win on Tuesday I think we are in for sure, but if we lose I think we start watching scoreboards and counting spots and hope for some help.
I glanced again at the 50webs site this morning. But instead of looking at all 79 of the brackets they feature, I focused only on the brackets that were made the past two days, and thus included our loss to Seton Hall in our resume. Of the 36 brackets, Marquette was still listed as a tournament team on 33 of them. Our seed ranged from 7 to 12 with an average seed of 10.
If we lose to Providence, we'll definitely be on the bubble, but if we beat them, the current numbers indicate we should be safely in the field of 68.
Great analysis.
It's good to know that the only weak areas - relative to the "contenders" - of MU's resume are two criteria that are over-hyped (team RPI/team OOC RPI) and another whose magnitude on the selection process is unknown at worst & questionable at best (total losses).
The general consensus amongst the pundits/insiders (e.g. Lunardi-types) is that the committee likes to focus upon record vs. Top X, quality wins, & bad losses. There is also reasonable suspicion amongst the so-called "bracketologists" that Sagarin and Pomeroy are relied upon more heavily than anyone on the committee or at the NCAA cares to admit (of course, Marquette's 6-seed last year did nothing but substantiate said suspicion). Isolate these metrics, and Marquette is a lock compared to its bubble-brethren.
The other metric which I'm still trying to get a grasp on is "Record vs. the Field." IIRC, during the interview portion of the Selection Show about 3 or 4 years ago, one-half of Gumbel2Gumbel was talking with the committee chair about one of the bubble teams (I forget who) that was included, and the chairman said that one distinguishing factor was how Team X fared against the other 64 teams already in (or projected to be in) the tournament. I don't recall seeing this metric mentioned otherwise by the committee or media mock-committee since.
My guess is that "Record vs. Field" is a tie-breaker, i.e. given similar resumes between the last two teams for one at-large slot, preference is given to the team that performed better against the already-selected field. In this respect, I would think that Marquette would be in a solid position, especially since there's a reasonable chance that Gonzaga, Bucknell & UW-M will all be in the tourney before Selection Saturday when the bubble discussion really gets intense. In fact, MU could end up with only 1 loss to teams outside the Field - SHU... compared to the 11 teams ahead of MU on the chart, only Georgia and Michigan St. could match this feat - no one can beat it.
I agree that everything lines up well for Marquette and beating Providence at the garden should seal the deal. A loss to PC hurts, but would not by any means be devastating.
The quality of this post is outstanding. I feel like its info you'd only get on a premium paid subscription site.
Please remind me, and others, of this thread(and others like it) when we next are trying to raise donations for some worthy cause, as I will be shamed/reminded to gladly chip in, paying over $ I would have gladly paid for subscription cost instead to the chosen donation cause.
Thanks, so very much, for all the work put into intelligent, well researched posts.
One team that interests me greatly is Tennessee. Lunardi has them as a 9-seed, as does 50webs currently. However when you really look at their profile, they could either be much higher or a bit lower. If you go strictly by the computer rankings (kenpom and Sagarin), they're probably a 12 or 13 seed, barely making the field. But when you look at their RPI, SOS, and their plus/minus score, they could easily be a 7 or a 6.
I think they provide an excellent test of how much the computer rankings play in. If Tennessee is an 11 or 12 seed this year, it's safe to say that Sagarin and kenpom get a heavy influence, even if the NCAA SC denies it, especially relative to Marquette's seed last year. If they earn a 6 or 7, it probably indicates they still rely just as heavily on the RPI. Of course, if they're in the 8-10 range, then it's anyone's guess ;D
Quote from: houwarrior on March 07, 2011, 10:48:38 AM
The quality of this post is outstanding. I feel like its info you'd only get on a premium paid subscription site.
Please remind me, and others, of this thread(and others like it) when we next are trying to raise donations for some worthy cause, as I will be shamed/reminded to gladly chip in, paying over $ I would have gladly paid for subscription cost instead to the chosen donation cause.
Thanks, so very much, for all the work put into intelligent, well researched posts.
We don't do ads. We don't try and sell stuff. Just once a year we ask people to donate to Al's Run. Donations still accepted. Just sayin'
http://www.crackedsidewalks.com/2010/09/als-run-2010.html
http://www.firstgiving.com/MarquetteFans