Sunday, February 27, 2005

Sorting out the Big East

It used to be that the tournament committee would use conference record as a handy tool to separate bubble teams within the same conference. It made sense - the team that does better in a group of 16 to 18 common games has proven to be better.

But that was before we had mega-conferences and the imbalanced schedules that come with them. Of the power conferences, only the Pac-10 still employs a balanced schedule. In 2005, when it's time to evaluate bubble teams within a conference, it not only matters who was on a team's out-of-league schedule, but also who they played within their conference.

The Big East schedule uses an NFL-style parity system. For television purposes, good teams (as judged before the season) play other good teams more often than the bottom-feeders. With Notre Dame's loss to UCLA on Sunday, there are four teams in the conference whose at-large status is could be equally argued positively or negatively. All four Big East hopefuls could end up finishing deadlocked with a 9-7 conference record. But would those 9-7 marks all truly be equal?

Fortunately, the 12 team Big East breaks down nicely into three distinct groups, making this easy to assess.

Group A, "The Locks" - Boston College, Syracuse, Connecticut, Villanova
Group B, "The Bubble" - Notre Dame, Pittsburgh, Georgetown, West Virginia
Group C, "The Dregs" - Seton Hall, St. John's, Rutgers, Providence

Here's how each bubble team has fared against the respective groups, with future games in parentheses:

Team            A      B      C     Total 
Notre Dame     3-4    2-2(1) 3-0(1)  8-6 
Pittsburgh     3-2(1) 1-3(1) 4-1     8-6 
Georgetown     1-4(1) 3-1    4-1(1)  8-6 
West Virginia  0-5    2-2    6-0(1)  8-7 

Notre Dame leads the way with seven games against The Locks, and will end up playing the fewest against The Dregs. Their three wins against group A make their 8-6 record the most impressive on this list. Pitt equals the Irish's feat of three wins against The Locks, making them a close second.

We can get a crude estimate of a team's schedule strength by giving them two points for each game played against group A, and one point for each game against group B. I'll also show each team's non-conference RPI to get an equally crude estimate of their non-conference performance.

             Schedule
Team          points   NC-RPI
Notre Dame      19      162 
Pittsburgh      17       83 
Georgetown      16      123 
West Virginia   14       62 

This paints a very muddy Big East bubble picture. The team with the best in-conference performance fared the worst out of conference and vice versa.

Really, any of these teams that gets nine wins has a good case for a bid. But under the new RPI formula, which penalizes each of these squads anywhere from 12 to 28 spots, nine wins must also be combined with avoiding a bad loss at MSG to secure a bid.

Even with the loss to UCLA, Notre Dame is in the best position of these teams, and should be able to get in with a win at home against lowly Rutgers on Wednesday.

With an RPI dangling in the low 60s, Pitt needs a split in the last two regular season games (at BC, at Notre Dame). Pitt's policy of bullying weak teams at home in November and December will not be popular among committee members.

The most likely scenario for Georgetown is that they lose at UConn on Wednesday and win Saturday against Providence at MCI. Georgetown doesn't get much help from their non-conference performance and their poor play down the stretch. How will the committee view the Hoyas win at Davidson in November? They deserve some credit for going there, but is that win better than Notre Dame's win at Indiana or West Virginia's win at NC State? According to the RPI it is.

West Virginia must win at Seton Hall on Saturday to join the nine-win club or they will have to win the Big East tourney. At 9-7, they would probably get the nod over Georgetown based on recent play and the wins over LSU, GW and NC State.

Never before have eight teams advanced to the NCAA tournament from one conference. Usually, it has as much to with the fact that one bubble team's success comes at the expense of another when they play each other. That's not the case in the Big East this season.  There is only one game between these teams this week.

In the Big East tournament, three of these teams will draw a member of The Dregs in the first round. Should all three advance, then the four bubble teams would each square off with a member of The Locks in the quarterfinals, avoiding any confrontation until the semis. It will provide a worthy final exam, and one that could possibly be passed by all.

Thursday, February 24, 2005

Block Percentage

I'll admit it. I am in love with blocked shots. Dominating shot blockers are rare, and teams that have one seem to be very good defensively as a team. But this was just my subjective opinion until today, when we can look at (on the stats page) how often each of the 330 D1 teams rejects the opposition.

block percentage = (blocked shots) / (opponents' field goal attempts)

Block percentage is the percentage of an opponent's shots a team blocks. Sure enough, this stat correlates well to a good defense - better than turnover percentage, anyway. Among the top 30 in block percentage are seven of the top 15 defenses in the nation. None of the top 15 defenses can be found in the bottom 129 of this stat. Not surprisingly, UConn dominates this metric, sending back one shot for every seven the opposition attempts.

As usual, there are some anomalies:

             Adj Def   Block %
Team          Eff Rk    Rank 
UNC              1      117 
Washington St.   2      169 
Oklahoma St.     9      201 
Colorado       183       17 
Iona           177       13 
Fairfield      119        2 

As a team in a power conference, you can have a great defense without blocking a lot of shots, but it's almost impossible to block a lot of shots and not have a great defense. Colorado is the only power (top six) conference member in the top 90 in block percentage but in the bottom half of the adjusted defensive ratings.

Comments?

Wacked Out Wednesday

I've had a love-hate relationship with West Virginia all season. More like a love-ignore relationship. I forecasted good things for them before the season. I founded the D'Or Fischer fan club after the 10-0 start. Then I sheepishly had to abandon the Mountaineers (and the game diary concept) after it became apparent they were a team that relied on the jump shot, but didn't have enough guys who could make one.

The Mountaineers came back from the dead on Wednesday - both in the game they played and for the season. They overcame a 14-point deficit with ten minutes left to win at #18 Pittsburgh. Simultaneously, nearly every other bubble team in action lost - and a few of them suffered crippling defeats that should forever take them out of the at-large discussion.

Miami of Ohio, Georgia Tech, Old Dominion, South Carolina, Vanderbilt, George Washington, Memphis, and Iowa - all teams with at-large aspirations Wednesday morning - were losers. Only the first two on that list have serious thoughts of an at-large bid now.

And now so does West Virginia, even with a loss to 5-19 Marshall blighting the tournament application. WVU, winners of five of their last six, finishes at home against Rutgers (a nine point winner over Arkansas Monticello last night) and at Seton Hall. They should be favored in both. If they win those two, a 9-7 Big East record with non-conference road wins over LSU and NC State should be enough.

With 6'11" marksman Kevin Pittsnogle on fire (interesting coincidence: Pittsnogle is Latin for "one who leads his team to victories over Pitt." He scored 49 points in the season sweep of the Panthers), maybe WVU can salvage a bid.