Sunday, February 29, 2004

Bubble Analysis

Two weeks until the speculation can end. Conference tournaments will decide the fate of most teams that still have questions about their at-large status.

(One question mark means you're in, but shaky. Two question marks means you're out, but can get in with a strong finish.)

ACC (6 teams in) – Duke, North Carolina St., Georgia Tech, Wake Forest, North Carolina, Maryland(?), Florida State(??).

Plenty of folks think 6-10 in ACC play is good enough to receive a bid. My rule of thumb is any team has to be within a game of .500 to earn an at-large bid.

This statement can be found in Michael Wilbon’s column in Sunday’s Washington Post. I love how he applies his own ‘rule of thumb’ to the selection process. How about this: let’s apply the committee’s rules and go from there? Here are the only two rules the committee is given when picking teams…

I. Principles for Selecting At-Large Teams

1. The committee shall select the best available teams to fill the at-large berths, regardless of conference affiliation.

2. There is no limit on the number of teams the committee may select from one conference.

I think Maryland has a strong case as being one of the ‘best available teams.’ They are 14-11 overall, 2-9 against the top 25. It’s obvious that they aren’t a top 25 team so let’s ignore those games. They are 12-2 on the rest of their schedule. Their losses are at #46 and an almost-home game against #91. Their quality wins include home vs. #47, @#52, and a home/road sweep of #75, and a home win vs. #94. They still must get a win in the finale hosting #52 UVa. Assuming they do, they would be 5-1 against teams ranked between 26-75. Considering the final at-large team will probably get a 12 seed, one needs to be in the theoretical top 45 or so to get a bid. With a solid record against teams in this range, and two quality non-conference wins (at Florida, home vs. Wisconsin), I’d put Maryland in if they split their last two. A 1st round loss in the ACC tourney to Georgia Tech would complicate things and may leave them at home.

Florida State on the other hand is in dire straits. They need a win in the finale at Georgia Tech. Wilbon, among other casual observers, believes Florida State has a better chance than the Terps of getting in with a 6-10 record. This is probably based on FSU’s higher win total. They picked up 2 extra D-1 wins by virtue of participating in the prestigious “Pittsburgh Holiday Classic.” Their two wins were against Wagner and Chicago State. Does this make them better than Maryland? Obviously not. They don’t have a quality non-conference win. Their record against the top 25 is 3-8, and against 25-75 it’s 3-3. 6-10 in the ACC won’t get it done for the Seminoles unless they get to the ACC championship.

SEC (7) – Mississippi St., Kentucky, Florida, South Carolina, Alabama, LSU, Vanderbilt, Georgia(??).

Georgia is still lacking a good road resume. They have two wins away from Athens: at Kentucky which is worth a lot, and against Clemson in Atlanta's Philips Arena. They also took Gonzaga into overtime in Spokane which may be taken into consideration. A win at Florida or a trip to the SEC semis would make them hard to leave out.

Big East (6) – Pittsburgh, Connecticut, Providence, Seton Hall, Boston College, Syracuse, Notre Dame(??), Rutgers(??).

Big 12 (6) – Oklahoma St., Texas, Kansas, Missouri, Oklahoma(?), Texas Tech(?), Colorado(??).

The Big 12 tourney is more intriguing than any other conference tourney this year. Suddenly, Wednesday’s Missouri @ Texas Tech game has become more important for the Red Raiders than the Tigers. Tech has lost 7 of their last 10.

Big 10 (3) – Wisconsin, Michigan State, Illinois, Michigan(??).

Conference USA (6) – Memphis, Cincinnati, Charlotte, Louisville, UAB, DePaul.

Atlantic 10 (3) – St Joseph’s, Dayton, Richmond(?).

Richmond needs to make noise in the A-10 tourney to get the committee’s attention again. The fact that the A-10 decided to let St. Bonaventure play in the tourney was not good news, because Richmond would have otherwise had a 1st round bye as the 3rd seed in the A-10 West.

Pac 10 (2) – Stanford, Arizona.

Arizona barely hangs on to safe status this week.

Mountain West (3) – Air Force, BYU, Utah.

All of these teams can play their way out.

Air Force has no room for error, and must win their 2 games this week and then beat Colorado St. in the 1st round of the MWC tourney.

The winner of Utah at BYU on Big Monday makes a much stronger case, especially if the winner is Utah. While BYU is on a 6 game win streak, 4 of the 6 games were at home and the two road wins were against the two worst teams in the MWC. I am not impressed.

WAC (2) – Nevada (?), UTEP(?).

Both of these teams have to go deep into the WAC tourney to have any hope.

----------------------

44 bids are available from these 10 conferences, and I listed 50 teams. So that leaves 6 teams on the outside looking in. Teams from other conferences getting consideration: Gonzaga (a lock from the WCC), Southern Illinois (a lock from the MVC), Utah St. (looking good from the Big West), and Western Michigan(MAC).

So here’s this week’s Doubting Dozen. The top six teams would be in, the bottom six are out. Gonzaga, Southern Illinois, and Utah State are assumed to get the automatic bid from their conference. But if they don’t, the bubble divider gets moved up one team. The same applies if a team not listed from the 50 teams above gets one of their conference’s automatic bids. That means that while Richmond is listed as in, they would only get in under the ideal and unrealistic scenario that nothing stupid happens in conference tournaments.

1) Texas Tech
2) Maryland
3) Oklahoma
(Western Michigan)
4) Nevada
5) UTEP
6) Richmond

---bubble divider---

7) Georgia
8) Florida St.
9) Colorado
10) Michigan
11) Notre Dame
12) Rutgers

Thursday, February 26, 2004

Rebounding 2’s vs. 3’s

Here’s what I have come up for offensive rebounding percentages (OR%) for each type of missed shot. This is based on all games involving at least one D-1 team played between January 1 and February 21.

3 point shot: 21.0% (21% of missed 3 pointers are rebounded by the offense)
2 point shot: 41.1%
Free throw: 20.3%

Keep in mind these are NCAA-wide averages. Individual teams will be better or worse than these percentages depending on their personnel and who they play against.

If you question the validity of the numbers (which you should), just watch a few games with this in mind and see if it’s believable.

I was going to post a bunch of statistical mumbo-jumbo on the meaning of this, but I will refrain for now.

The one thing I’ll point out is this: The results give the 3 point shot an interesting symmetry. They’re worth 50% more than the made 2 pointer, but then they’re also about ½ as likely to be rebounded by the offense. This makes it a little easier to understand shot selection. If your team’s brickmeister is jacking up a lot of perimeter shots, he becomes a drain on the offense. Sure, there are nights when he’s hitting and everything’s great. But over the long haul he’s going to miss too often to help the offense. However, the closer he gets to the hoop, the less his bricks hurt, because there's a better chance points will be scored on a put-back. You can also look at it from the defensive perspective in the same light.

For those with too much time on their hands, the methodology for how I came up with the percentages is described below.

----------------------------------------------------------

From the box score we know the following:

A team’s missed free throws
A team’s missed 2 point FGA’s
A team’s missed 3 point FGA’s
A team’s offensive rebounds*
Opposing team’s defensive rebounds*

*Team rebounds – the rebounds that go out of bounds before being controlled by a player - throw a monkey wrench into this. In the “official box score,” team rebounds are recorded as both offensive and defensive. Unfortunately SportsTicker, the purveyor of college basketball box scores, dumbs down the official box score and lumps team rebounds into one category. So for every game, I made the assumption that team rebounds go to the offense as often as normal rebounds (about 1/3 of the time).

Things we can’t tell from the box score:

1) How many offensive rebounds occurred on FT’s, 2’s or 3’s.

2) How many FT misses had rebounds – some misses will occur on the front end of a 2 or 3 shot foul.

I am trying to solve for (1). To get around (2), I took Team A’s offensive rebounds + Team B’s defensive rebounds and subtracted the missed 2’s and 3’s by Team A. The remaining number yields the number of missed free throws that were rebounded. In some cases, due to the team rebound assumption, the number of remaining rebounds was greater than the number of missed free throws. Corrections were made for this.

So now to figure out what’s mentioned in (1). I assumed a 2 pt OR% and a FT OR% for all games. The 3 pt OR% can be calculated from the total offensive and defensive rebounds for all games after the first two are assumed.

With the assumed values I went back and computed the theoretical rebounds for each team in each game. This was done taking 2 pt. OR% x 2 pt. misses + 3 pt. OR% x 3 pt. misses + FT OR% x “reboundable” FT misses. From this I took the difference between the theoretical rebounds and actual rebounds and called this the rebounding error. I averaged this error over all games. The numbers presented at the beginning of this post represent the values of OR% for each shot that minimized the rebounding error. If you have any more questions, e-mail me. I could go on, but I think I’ve covered the main points here.

Wednesday, February 25, 2004

Bursting Day

Kent State lost at Buffalo 82-68 tonight, a team that had no previous top 100 wins. There's no way to give Kent any benefit of the doubt here, they got waxed from start to finish by a bad ballclub. This effectively ends the Golden Flashes at-large chances, barring possibly a run to the MAC championship against Western Michigan.

Richmond got knocked off by Rhode Island, and now probably needs a win over St. Joe's in the A-10 tourney to get in.

Elsewhere, Florida State missed a golden opportunity to sew up a bid by losing at Wake Forest in overtime. They blew a 15 point lead with 10 minutes to play.

The one team the helped itself was DePaul, getting the road win at Louisville. CUSA has been a sharing conference this year, with good RPI teams Louisville and Cincinnati freely giving away games to the rest of the league. This is a big reason they may be a 6-bid league this year, with nobody getting a top 4 seed.