Thursday, March 03, 2011

W-1.5 BP68

Conference tournaments are under way and March Madness is definitely here. Please check out the conference tournament previews that are already posted here. I will post more tomorrow, and will continue posting them over the next week. So please keep checking back.

The next BP68 will be out after Saturday night's games. I will also begin my daily bubble watches on Friday.

Remember, this is a projection of where things will end up and not a snapshot of where things are now. If your team is not rated where you think they should be, please tell me what you think I did wrong in the comments and I'll be happy to discuss it with you. One other note is that teams in CAPS are teams I'm projecting to earn automatic bids.

As always, here is how I see things ending up:

1. KANSAS (BIG 12)

2. Purdue
2. BYU (MWC)
2. Notre Dame
2. Syracuse

3. Texas
3. San Diego State
3. Wisconsin
3. North Carolina

4. Louisville
4. Florida
4. UConn

5. Cincinnati
5. Arizona
5. St. John's
5. Georgetown

6. Villanova
6. West Virginia

7. Vanderbilt
7. Texas A&M
7. Xavier

8. Missouri
8. Tennessee
8. Kansas State
8. Michigan State

9. George Mason
9. Illinois
9. Georgia

10. Richmond
10. Virginia Tech

11. Marquette
11. Florida State

12. Baylor
12. Boston College
12. Michigan
12. Gonzaga
12. Maryland



15. LIU (NEC)


Teams seriously considered that just missed the cut:
Clemson, Colorado, Nebraska, Memphis, Missouri State, Alabama

Decent resumes, but not good enough:
Penn State, VCU, Southern Miss, Cleveland State, Colorado State, New Mexico, Washington State

Long shots, but still in the at-large discussion:
Miami (Fl), Dayton, Duquesne, Minnesota, Oklahoma State, Drexel, Hofstra, Marshall, UTEP, Princeton

Still alive, but pretty much need a miracle:
Virginia, Northwestern, James Madison, Tulsa, Valparaiso, California, USC, Arkansas, Mississippi, Mississippi State


Anonymous said...

Maryland?! MARYLAND?! The same Maryland team that has an ELO_Chess of 64, an RPI of 88, is 7-8 in the 5th rated conference, has an amazing 0 Top 50 wins and is 4-12 vs the Top 100, has a 4-9 road/neutral record, and is 5-5 in their last 10 games? That Maryland? I don't care how soft the bubble is, Maryland isn't even on it right now.

I know you aren't stupid enough to think they're in the field right now with those numbers, so I guess you think they can make up ground because they have a top 20 predictor rating. News flash: there's no time left for them to pick up enough big wins. They have 1 game left in the regular season, against a sub-100 Virginia team. In addition, they'll play at most 2 top-50 teams and 1 top-25 team before the ACC final. That means, if they made it to the finals and lost, they'd have an ELO_Chess in the high 40s-low 50s, a 2-10 record against the RPI top 50, and a 6-13 record against the RPI top 100. Even in this weak year, that's not going to be good enough.

Maryland has zero chance at an at-large bid after losing tonight. Their only path to the NCAA tournament is the ACC auto-bid.

Jeff said...

For one thing, road/neutral and "last 10" really doesn't matter. Second, the lack of big wins is pretty commmon on the bubble. Have you looked at the resumes of UAB or Memphis or Gonzaga or Clemson.

I think Maryland has a great path in the ACC tournament, and should make the semifinals, where they'll have a 50/50 shot of making the finals. Right now that puts them ahead of the teams I've got out of the field.

Who exactly is it with an awesome resume that should be in? Cleveland State??

Anonymous said...

Agree with the first comment. Maryland has to win the ACC tourney. I know they claim conferences don't matter, but Va Tech, BC, FSU, and Maryland are not all getting in. I think 2 of those 4 prolly.

Jeff said...

Conferences do not matter. And it's most certainly possible for all four of those teams to get in.

Anonymous said...

The committee claims conferences do not matter. ACC will get 4 teams max if Duke or UNC win the tourney.

Jeff said...

Yes, you are correct that some things the Committee claims it doesn't do, it in fact does. But I can tell you from experience that the Selection Committee does not count conference bids. In no way is it assured or even likely that the ACC will be limited to four bids if UNC or Duke wins the ACC tournament.

Jeff said...

For example, there's a chance the Big Ten will get only four teams into the Tournament even though it's probably the best conference top to bottom. The Selection Committee isn't going to give Michigan or Michigan State a pass on a weak resume just because their conference is strong.

Teams are judged on their own.

Anonymous said...

We'll just disagree I guess. The BE would not be getting 11 bids if it wasn't the #1 RPI conference. Sagarin has the BE #1 too, so it's interesting you think the Big 10 is the best.

We'll find out next weekend, but those ACC teams will be playing elimination games against each other in the ACC tourney.

Anonymous said...

Vandy a 7? Not a chance. If they lose at home to Floriday, they might slip to a high 6 but no way are they lower than 6.

Jeff said...

Actually, on pure average top to bottom, Sagarin has the Big East and Big Ten even. Sagarin has the Big East higher than the Big Ten on "central mean", which in that case cancels somewhat out how awful the bottom of the Big East is (the strength of the Big Ten is how amazingly strong the bottom of the conference is - there are no bad teams)

I would also add that Sagarin's conference ratings take his "synthesis" team ratings, which are a combination of the ELO_CHESS and PREDICTOR. If you want to know which team is "best", you look at the PREDICTOR, and Sagarin's PREDICTOR says that the Big Ten is the best conference, which agrees with Pomeroy.

And no, the fact that the Big East is the #1 conference is not why they're likely getting 11 teams. It's because there are 11 Big East teams that have resumes better than all of the teams that are going to be left out of the field. Which team do you think should be out? At this point, only Marquette is even debatable - the other 10 are clearly in.

Anonymous said...

Just went to Sagarin's website and the BE is higher in all 3 of his ratings he uses for conferences. Can you please give a link to your info?

Jeff said...

If you take a straight average of the Sagarin ratings, for the Big East it is 84.35, and for the Big Ten it is 84.26. Basically a wash. Those numbers should be on whatever Sagarin website you're looking at.

But as I said, those ratings are based on the "synthesis" rating that Sagarin does. If you average up the PREDICTOR ratings (you can do this by hand with a calculator and about 60 seconds of free time) you get an average of:

Big Ten: 84.83
Big East: 84.01

Not that it matters. As I said, the Big Ten doesn't get to have more teams because it's rated higher as a conference. It's simply not the way the Selection Committee works. I can give you many examples of Conference A being rated better than Conference B and getting fewer teams.

Anonymous said...

"Actually, on pure average top to bottom, Sagarin has the Big 10 and Big East even." This is false then. What's more troubling is you seem to have known this and disregarded it because it did not back your argument.

Jeff said...

How was that false? They are even to within a tiny fraction of the margin of error.

Computer ratings are not perfect. Understand them and use them as they are meant to be used, and don't try to draw any conclusions beyond that.

Anonymous said...

EVERY year we hear how tough the BE is, but if you a results person, not so fast my friend. In the last 20 years tourney appearance in the final 4 is as follows; BE= 9, ACC=19, Big 10=15, SEC=17, B12=7. That too far back? Think they are on a late surge? The last 6 yrs? BE= 5 final fours, ACC=4, B10=5, SEC=3, and B12=1. The BE is constantly hyped by ESPN (the homey network) , but in the tournament they fail to perform particularly given that they get an extra 3-4 teams in to start. Vandy loses and they drop 3 or 4 seeds, Villanova goes on a putrid run and it gets shrugged off to the "toughness" of the big east. Pure marketing my friends. ESPN knows where the population is and wants to get it fired up to watch.

Jeff said...

You're preaching to the choir. I've talked many, many times over the past few years about how over-hyped the Big East is.

Anonymous said...

20 years is a very small sample. Besides that, number of final fours is a poor metric. What if a conference has the 5th through 15th best team? Maybe no final fours but a hello of a conference. I'm an ACC fan, but there is little doubt the BE is the best conference this year. 11 teams in the tourney is quite a feat.

Anonymous said...

The BE has also only had 16 teams for 4-5 years, so any comparison stats should only include that timeframe. Most would agree it's a completely different conference than even 10 years ago.

Jeff said...

Actually, there's little doubt that the Big Ten is the best conference this year. Don't look at the number of Tournament teams - that's an arbitrary metric. The top three Big Ten teams are as good as any team the Big East has, and look at the bottom of the conference. There are no bad teams. The last conference to have zero bad teams was last year's ACC, actually, but the top of the conference wasn't nearly as strong as this year's Big Ten. A better comparison would probably be the 2006-07 ACC.

It is correct that Final Fours is another arbitrary metric. A better judge is "wins vs expected wins", which compares how well teams have done relative to what teams with their same seed have averaged. This takes into account all teams in a conference, and acknowledges that a 12 seed making the Sweet 16 is impressive and a conference deserves credit for it. By those metrics, the Big Ten has been the best performing conference for the past 20 years. The Big East has underperformed its seeds on average.

Anonymous said...

I prefer to use medians when conducting analysis. A median is the point that half of a population will fall below and half will fall above. I think it's a decent measure of a conference's relative strength. Which conference has the greatest median ranking according to various rating systems (Pomeroy, Sagarin, RPI, Massey, etc)?

Anonymous said...

My look at just the last 5/6 years indicates the BE is less successful reaching the semis than the B10,ACC especially when you factor in the fact that ESPN hypes in 3,4 extra teams each year.

Anonymous said...

Since the BE expanded in 2006, the conferences have been represented as follows in the final 4:

BE: 5
ACC: 3
Big 10: 2
SEC: 3
Big 12: 1
Pack 10: 3
CAA: 1
Horizon: 1

Anonymous said...

B10 has 4 bad teams. The BE has 5 bad teams. 5/16 is much better than 4/11.

The B10 has 2 really bad teams, the BE has 2 really bad teams.

It's not even a comparison anymore, BE is clearly the top conference.

Jeff said...

How do you define "bad" and "really bad"? Here are the bottom 4 teams in the Big Ten vs the bottom 5 teams in the Big East, with their Sagarin PREDICTOR/Pomeroy ratings:

Big Ten:
Penn State 61/55
Northwestern 65/64
Indiana 70/71
Iowa 84/90

Big East:
Seton Hall 63/62
Rutgers 81/81
Providence 83/86
South Florida 115/127
DePaul 185/185

The well accepted two best measures of team quality are the Sagarin PREDICTOR and Pomeroy ratings. If you take the average of those ratings for all teams in the conference, both ratings place the Big Ten as clearly better.

If you want to define conference quality by some other metric than the average quality of the teams then that is fine. Sagarin likes to take a weighted average, which minimizes the impact of one really bad team, and that's fine. But even in the case of Sagarin, you've got to do some math yourself, because his conference ratings are a combination of his ELO_CHESS and PREDICTOR ratings, and the ELO_CHESS is a measure of resume quality, not team quality.

Anonymous said...

The mean (simple average) is a poor measure of the central tendency, especially when a distribution is undefined. A much better measure is the median. Which conference has the highest median Sagarin predictor and pomeroy rating?

Jeff said...

Actually, median would be a terrible way to do it. It means that the quality of only one team matters, and a conference is not punished for having some terrible teams.

You can't underrate how much it changes a conference when there are zero easy games. Big East teams know that they can sleepwalk through DePaul. ACC teams can sleepwalk through Wake Forest. There are no easy games in the Big Ten.

I do agree that a pure mean is not perfect. A weighted mean is probably more ideal, so that it discounts the impact of one team on the ratings.

If you want to calculate the weighted mean of the two conferences you can. I'm not in a location right now where I have the right computer software to do that... maybe if I have some free time tonight.