Sunday, October 9, 2016

Offseason bracketology 16-17

The 1 line: Duke, Kansas, Villanova, Oregon
The 2 line: Kentucky, Wisconsin, North Carolina, Virginia
The 3 line: Xavier, Michigan St, Arizona, Iowa St
The 4 line: Louisville, Indiana, California, Dayton
The 5 line: Purdue, UConn, Texas A&M, West Virginia
The 6 line: Baylor, Rhode Island, Marquette, Virginia Tech
The 7 line: Gonzaga, UCLA, Florida, Creighton
The 8 line: Michigan, Miami, Wichita St, Maryland
The 9 line: Colorado, Auburn, Cincinnati, Monmouth
The 10 line: Texas, St Mary's, Notre Dame, VCU
The 11 line: Florida St, USC, Seton Hall, Oklahoma, Davidson
The 12 line: New Mexico, Illinois St, Valparaiso, Harvard, San Diego St
The 13 line: Akron, Towson, Long Beach St, North Dakota St
The 14 line: Chattanooga, Texas-Arlington, Belmont, UAB
The 15 line: Weber St, Lehigh, Vermont, North Florida
The 16 line: Winthrop, Sam Houston St, Fairleigh Dickinson, New Mexico St, South Carolina St, Texas Southern

Next 4 in:
Florida St
Seton Hall

Last 4 in:
New Mexico
Illinois St

Last 4 out:
Northern Iowa

Next 4 out:
Ohio St

Break it down!
Big 12 6
Big 10 6
Pac-12 6
Big East 5
A-10 4

Wednesday, May 4, 2016

The RPI sucks

It's time for another rant against the RPI.  I've said the things I'm about to say before.  In fact, I encourage you to go back to last April's archives and find the RPI/SoS post there.  I won't reprise the mathematics here, but I'll refer to the end lesson.  The lesson is this:  in building a schedule, the current system emphasizes bad team avoidance more than playing good teams.  It rewards teams more for simply playing below-average teams instead of the bottom of D1, instead of rewarding teams who actually play good teams.

It's fair to say the overall system is biased against mid-majors.  Note that you have to be careful in what you mean about bias, though.  It's not that very few mid-majors get at-large bids.  It's that the system is set up in a way to make earning an at-large bid impossible for mid-majors.  The current system, based on how the RPI is calculated, favors teams who have bad team avoidance, which is something by definition that the mid-majors cannot do.  There is no recourse for a mid-major to prevent this fatal flaw from ruining their resume.

Knowing this, let's brainstorm a new version of the RPI.  This new version should still punish teams who willingly schedule many cupcakes, but it should reward teams who schedule up and attempt to get quality wins, without erasing those gains unfairly.  There's two main ways to do this:

1) Throwing out outliers.  You know in some judging competitions, how the high score and low score get thrown out when there are many judges?  Let's introduce that to the RPI.  For the SoS calculation, throw out the best team and worst team, and then calculate the SoS.  Or even just throw out teams on one end.  Throw out the worst 3 records in a team's non-conference schedule before calculating the non-con SoS.  In fact, this is my favorite solution.

What does this do?  By throwing out the bottom 3 teams on a team's schedule when calculating non-con SoS, you allow everyone a couple of free cupcakes without hurting their computer numbers, which is okay.  You keep the truly horrendous teams from causing extra damage to a good team's profile.  If a team still has several other cupcakes on the schedule, now they're still going to be punished, even more severely when compared to teams with just a couple of cupcakes.  And finally, with less teams factoring into the SoS calculation, the excellent teams will stand out more clearly.

2) Artificial scaling.  For the SoS calculation, put in a floor.  If an opponent has less than a .250 record...bump it up to .250 for the SoS calculation.  Or .300.  Or .200.  Or any number you want.  This would require some research that I don't have time for, but I'm sure an optimal number is out there.  Or you can introduce a sliding scale, where all below .500 records receive a proportional bump in value.  A .200 record becomes .300, a .400 record becomes .450, you get the idea.

What does this do?  It softens the blow that a truly terrible team has on the SoS.  It mitigates the bad team avoidance issue by helping prevent a single bad team from having excessive influence on the SoS.

I'd like to see someone figure out a system incorporating one or both of these ideas.  We need something to lessen the impact of cupcake games on the SoS.

Monday, March 28, 2016

Random musings

I'm working on a post where I'm going to propose some changes to the RPI, in order to come up with a metric that can be used to sort teams that isn't mathematically flawed the way RPI is.  But for now, I wanted to hit a few topics that, by themselves, don't warrant full blog posts.  But I want to get them out for public consumption.

1) Three times in 11 years, a bubble team makes the final four (Mason, VCU, Syracuse).  This just proves that bubble talk is actually important.  It actually matters who gets in and who gets out.  Syracuse was a true bubble team this year (although the media narrative that they should've been out is a bit too revisionist - they had a reasonable case).  Do we know that Valpo or SDSU can't make that type of run, especially with that region being cleaned out in the bottom half?  This is why bubble discussion matters, because there are several teams that didn't make the tourney that could've made the run Syracuse made.

2) This is going to sound blasphemous and counter to everything I stand for...but I think the NIT is going to have to get rid of autobids (or increase their field).  There's good teams being left out of the NIT, and there's a big imbalance between the 8 line and the 6 line in that tournament.  Now that the CIT is established, those low-level conference champs do have another home, and actually have a chance to play for several games instead of just 1, maybe 2.  It runs counter to my general philosophy, but I think this is best for everyone.

3) Vegas should keep their tourney at 8 going forward, and frankly I think all those teams should migrate back to the CIT anyways.  If anything, the CIT should adopt the Vegas 8.  Have 32 teams play down to 8 the first weekend, then send those 8 to Vegas.  Something like that would work best for the low-level tourneys.

4) It's important to seed the bottom four lines correctly.  This is what happens when you match a 1 seed (Michigan St) with a 14 seed (MTSU), only you call it a 2-15.

5) I know you shouldn't use tournament results to justify selection/seeding, but you can't tell me that mass ejection by the Pac-12 on the first weekend had nothing to do with their overvaluing.  More about this in the RPI post coming in the next couple of weeks.

Friday, March 18, 2016

Bracketing fixes

Ok, we have the committee seed list 1-68.  How should the tournament have been bracketed?  Guess what:  the committee messed a couple things up.  Here's what I got when I applied their principles to their S-Curve.

*note:  I didn't have immediate access to previous tourney matchups, to see if that blocked anything

@Des Moines
1) Kansas vs. 16) Austin Peay
8) Colorado vs. 9) Butler
4) California vs. 13) Hawaii
5) Maryland vs. 12) South Dakota St
3) Miami vs. 14) Buffalo
6) Arizona vs. 11) Vanderbilt/Tulsa
@St Louis
2) Xavier vs. 15) Middle Tennessee
7) Wisconsin vs. 10) Syracuse

1) Oregon vs. 16) Southern/Holy Cross
8) Texas Tech vs. 9) UConn
4) Duke vs. 13) UNC-Wilmington
5) Baylor vs. 12) Yale
@Oklahoma City
3) Texas A&M vs. 14) Green Bay
6) Texas vs. 11) Northern Iowa
@Oklahoma City
2) Oklahoma vs. 15) Cal St-Bakersfield
7) Oregon St vs. 10) VCU

1) Virginia vs. 16) Hampton
8) USC vs. 9) Providence
4) Iowa St vs. 13) Iona
5) Purdue vs. 12) Arkansas-Little Rock
3) Utah vs. 14) Fresno St
6) Seton Hall vs. 11) Gonzaga
@St Louis
2) Michigan St vs. 15) Weber St
7) Dayton vs. 10) Temple

1) North Carolina vs. 16) FGCU/Fairleigh Dickinson
8) St Joseph's vs. 9) Cincinnati
@Des Moines
4) Kentucky vs. 13) Stony Brook
5) Indiana vs. 12) Chattanooga
3) West Virginia vs. 14) Stephen F Austin
6) Notre Dame vs. 11) Michigan/Wichita St
2) Villanova vs. 15) UNC-Asheville
7) Iowa vs. 10) Pittsburgh

The 1 line:  no changes
The 2 line:  Villanova and Xavier should switch.  Not only does it reward Nova and keep them home, it also gives Xavier a better site to my eyes as well.  Instead of giving geographical advantages to the 7th and 8th teams, they choose to save the teams ranked above them.  I'm not sure that's how it's supposed to work.
The 3 line:  I agree.  I think OKC is a better geographical site for Miami, barely.  However, the chain reaction would've sent A&M away from OKC, and the difference between OKC and Provi for Miami is negligible, so it's a good switch to make.
The 4 line:  You could make an argument that Iowa St (Denver) and Kentucky (Des Moines) should switch regional sites.  Obviously ISU wants Des Moines, and Kentucky can't travel by car/bus either way.  It wouldn't be the worst change in the world to make.
The 5 line:  On first glance, I've got Maryland and Baylor flipping spots.  Obviously Maryland being ranked higher, they should stay east while Baylor goes to Spokane.  But the devil is in the details.  Baylor and Kansas played 3 times this year; therefore they can't be in the top half of the South bracket with Kansas.  This costs Maryland a cross-country trip.  Tough break.
The 6 line:  Here's a place where I think the committee could have gotten creative if it wanted to.  Currently, they have Arizona making a cross-country trip and Texas lining up against A&M.  If you send Texas to Denver instead, you can fit Arizona into OKC.  More fair to Arizona, less fair to Texas.  But, as I was working up this plan, Seton Hall gummed the whole thing up.  Seton Hall as a 6 played Xavier and Villanova three times each this year, locking them out of their two regionals.  Therefore, SHU has to play in either Denver or OKC, which messes up the whole Arizona/Texas plan.  Therefore, I think the committee got the 6 line right.  Arizona ends up with the short straw in this scenario.
The 7 line:  no changes
The 8 line:  Here's one of the stupider things the committee did.  Texas Tech has two options:  Raleigh and Spokane.  Both aren't great.  They're about equal.  However, if you send TTU to Raleigh, then you eventually have to send St Joe's to Spokane.  That's what this committee did.  Why not send TTU to Spokane and save St Joe's a couple thousand miles or so?  That's a sloppy mistake by the committee.  By the way, this also changes USC's fate just a bit, as they're in UVa's path instead of UNC's.
The 9 line:  I'm not sure anyone's in the spot they're supposed to be.  Providence as the highest on the 9 line should be in UVA's path, not UNCs.  A modest mistake here.  Butler is next and should get Des Moines.  UConn as the lowest should be the one of the four to bite the bullet and go to Spokane.  The committee made a pure mess of the 9 line, and I'm not sure why.  Why is Butler taking the trip to Spokane?  The only logic I see is UConn fans traveling better and the NCAA doing a blatant favor, but it's not like UConn is staying close to home - they're going to Des Moines.  This is weird.
The 10 line:  Shouldn't Pitt be in Brooklyn?  I know the committee put Temple there, and they're closer.  But Pitt is higher ranked and close enough to Brooklyn, compared to St Louis, to make it a real difference.  This change causes a chain reaction that changes 3 of the 4 7-10 matchups.
The 11 line:  The committee chose the right two sites to send PIG winners to.  The actual PIG matchups kinda feel arbitrary to me.  I'd have the best play the worst among the four, as there are no in-bracket conflicts that get in the way.  Wichita gets a minor, minor shaft by having to play Vandy instead of Michigan.
The 12 line:  no changes
The 13 line:  no changes
The 14 line:  no changes
The 15 line:  Middle Tennessee is the best 15 seed.  If they're going to play in St Louis (and they should), it should be against the worst 2, Xavier, not the best 2, Michigan St.  Middle Tennessee got a screw job here.  Everything else lines up fine.
The 16 line:  It sucks sending a winner from Dayton to Spokane, but the only alternative is moving the winner to the 15 line instead.  Nope.

Tuesday, March 15, 2016


I don't really feel like doing this, but whatever:

66 of 68 teams correct
34 seeded correctly
59 seeded within one line
2.56 average S-Curve miss

These are all marginal numbers.  66 and 34 seem to be above average with respect to everyone else; 59 isn't.  2.56 is also below the 2.25 target, but I'm far too lazy to calculate it for every other person out there.

7 teams I missed by more than 1 seed line:
Texas A&M (NCAA 3, Bracketball 5)
California (NCAA 4, Bracketball 6)
Oregon St (NCAA 7, Bracketball 9)
USC (NCAA 8, Bracketball 10)
Providence (NCAA 9, Bracketball 7)
Wichita St (NCAA 11, Bracketball 9)
Gonzaga (NCAA 11, Bracketball 9)

The committee went away from trends on the Gonzaga thing - usually conference champs get a seed bump in the process.  That applies to Wichita too.  However, on the other end, A&M got that bump.  So apparently that bump now applies to majors only?

The Pac-12 I underseeded as a whole.  I trusted the committee to look past the RPI gimmicks a bit; they didn't.  Oregon St is definitely a flagrant rank.  Did they see Cal's road/neutral record?  I don't like any of the Pac-12 seeds in this tournament, period.

Providence is probably the one mistake I think I made.  Too high on my end.

Other lessons:
1) Last year, I missed some 13-16 seeds when I tried to break down their resume too much.  The committee used the RPI and the conference RPI, and did a poor job seeding them.  So this year, I tried to mimick that strategy...and the committee went the other way.  PICK A WAY OF EVALUATING MID-MAJORS AND STICK WITH IT.  I'm very frustrated by some of these seeds.  Green Bay clearly got a name recognition bump with the Horizon...but UNC-Wilmington didn't despite the big year from the Colonial.  I give up on the 13-16 lines, there's horrible inconsistency.
2) The committee is very vulnerable to RPI manipulation.  Beware the conference that plays the game well; they'll get overseeded like the Pac-12 did.
3) Contingencies are meaningless.  The committee ignores Sunday results; they just do.

Monday, March 14, 2016

Vegas 16/CBI/CIT fields

This whole thing ended up being an adventure, but it looks like we finally have fields for all 3 of these shindigs.

Vegas 16:
Tennessee Tech vs. Old Dominion
UC-Santa Barbara vs. Northern Illinois
Oakland vs. Towson
Louisiana Tech vs. East Tennessee St

So, the Vegas 16 has 8.  HA.  Turns out all those high-majors they wanted...everyone turned them down.  Everyone.  I had speculated that this was the most likely tournament to attract the high-majors, and no one bit.  They had to go to 8 teams because they couldn't find more.  It's pretty damning for the future of this tournament, which I don't think is viable without those big boys.  Well, it could be viable if the CBI/CIT didn't exist, perhaps.  But otherwise, this might be a one-and-done.  They failed to get the types of teams they wanted.

As for the actual 8 teams?  Deserving.  It's not a bad field.  All 8 are certainly deserving of the postseason from a profile perspective.

Jackson St vs. Sam Houston St
South Carolina St vs. Grand Canyon
Mercer vs. Coastal Carolina
Louisiana-Monroe vs. Furman
Ball St vs. Tennessee Tech
UT-Arlington vs. Savannah St
Boston vs. Fordham
Norfolk St vs. Columbia
New Hampshire vs. Fairfield
Tennesse-Martin vs. Central Michigan
Army vs. NJIT
UC-Irvine vs. North Dakota
Texas A&M-Corpus Christi vs. Louisiana-Lafayette

Now, if you know how to can count to 26 teams.  Not 32.  The CIT could not fill a full field, and they don't need to look any further to Vegas, who siphoned off 8 teams who otherwise would've fit the CIT profile.  I would have thought, for all the world, that it would be the CBI that would suffer.

Since they didn't fill the full field, I can't be critical of any selection.  Army and NoDak, and maybe one or two of the HBCUs would've been marginal postseason picks in a normal year, but I can't criticize their inclusion in a year with only 26 teams.

Morehead St vs. Siena
Omaha vs. Duquesne
Western Carolina vs. Vermont
Albany vs. Ohio
Houston Baptist vs. UNC-Greensboro
Montana vs. Nevada
Pepperdine vs. Eastern Washington
Idaho vs. Seattle

Naturally, the CBI didn't get hurt while the other two got hurt.  I had predicted the exact opposite.  I really do wonder why teams preferred the CBI over the CIT.  Maybe these teams were expecting a few upper-majors to play.  Of course, none of them showed up.

Now, here, there's some questionable selections.  Seattle is 12-16 (7-7 WAC) with a RPI of 287.  Greensboro is 11-18 (10-8 SoCon) with a RPI of 237.  Western Carolina is 13-17 (10-8 SoCon), albeit with a tolerable RPI.  Duquesne is under .500 against D-1 competition.  There are some flagrant selections here.  Absolutely flagrant.

As it turns out, if you remove those 4 teams, plus 4 out of the back end of the CIT, and insert the Vegas have two pretty decent postseasons.  Instead, we have 3 marginal tournaments.

The lessons:
1) The majors have checked out.  No one simply wants to play 'em.  The Big 8 conferences don't, with the lone exceptions of Duquesne/Fordham, who don't get many chances.  The Mountain West had an avalanche of available teams, yet only Nevada accepted (and Nevada is on the rise and hasn't been a postseason team in awhile).  The MVC is skipping them all.  Seems like those 10 conferences, in general, are taking a collective stand.  That can be a problem.

2) I think the CIT may need either the Vegas to die, or to pare down to 16.  In the beginning, it was a great way for teams to finally make the postseason, as many of the small-majors had just their conference tournament every year as an option otherwise.  However, a few years in, with so many teams finally getting their feet wet, it's no longer a novelty.  Sure, there'll always be a few teams who've never played the postseason, or haven't played in awhile, but that list of candidates is slowly fading away as, year after year, a few teams leap off that list.  It's a simple case of attrition.

We'll see what happens going forward.


Presented, without comment:

Lowest RPIs to receive an at-large bid
72 Syracuse
63 Vanderbilt
58 Tulsa
57 Michigan
56 Butler
53 Pittsburgh
51 USC

Highest RPIs to miss the field
30 St Bonventure
34 Akron
38 St Mary's
39 Princeton
41 San Diego St
49 Valparaiso
52 Monmouth