Wednesday, March 14, 2007
Bracket Breakdown: South/East
I’ve been under the weather, so I’m not in the mood to write much. Fortunately, these two regions are not very complicated.
First, the South:
2nd Rd Sweet16 Elite8 Final4 Finals Champs 1 Ohio St. 98.62% 82.16% 68.80% 41.49% 21.00% 11.32% 3 Texas A&M 95.07% 70.87% 46.80% 27.35% 13.31% 6.91% 2 Memphis 98.03% 73.77% 34.47% 16.88% 6.59% 2.76% 6 Louisville 71.81% 23.52% 10.88% 4.33% 1.31% 0.43% 5 Tennessee 87.40% 51.42% 12.43% 3.29% 0.68% 0.15% 9 Xavier 61.50% 12.37% 6.84% 1.89% 0.41% 0.09% 10 Creighton 67.29% 20.26% 5.54% 1.68% 0.37% 0.09% 4 Virginia 89.61% 44.80% 9.40% 2.15% 0.38% 0.07% 8 BYU 38.50% 5.37% 2.36% 0.46% 0.07% 0.01% 11 Stanford 28.19% 4.89% 1.29% 0.29% 0.05% 0.01% 7 Nevada 32.71% 5.85% 0.94% 0.18% 0.02% 0.00% 12 LB St 12.60% 2.48% 0.12% 0.01% 0.00% 0.00% 14 Pennsylvania 4.93% 0.72% 0.08% 0.01% 0.00% 0.00% 13 Albany 10.39% 1.30% 0.04% 0.00% 0.00% 0.00% 16 C. Conn. 1.38% 0.09% 0.01% 0.00% 0.00% 0.00% 15 North Texas 1.97% 0.12% 0.00% 0.00% 0.00% 0.00%
This is the weakest region of the four, and just about every bracket will have Ohio State or Texas A&M. But Memphis and Louisville are legitimate darkhorse picks, especially considering Louisville’s semi-home court advantage in a 2nd round game against A&M, which isn’t accounted for here.
A while ago, I surmised that Ohio State could survive with Greg Oden on the bench better than the most pundits thought. But after doing the research, that was, um, wrong. Here’s Oden’s on/off plus/minus data for the ten most recent games that substitution data is available for.
On Off Min Michigan St. +10 -1 33:18 Michigan +17 -4 29:47 Purdue + 8 -1 21:08 Penn St. +11 -9 30:46 Penn St. +10 -2 22:47 Wisconsin + 7 -6 34:51 Michigan + 9 -5 32:38 Michigan +14 -4 27:30 Purdue +13 -2 32:28 Wisconsin +15 +2 23:01 Sum +114 -32 28:49 Oden On / 40 +15.8 Oden Off / 40 -11.5 Difference +27.3
Wow. Statistically, Ohio State is the most likely of the #1’s to go down in the first round. If I’m Central Connecticut, I go right at Oden every possession and flop when gets the ball in the post. Oden might end up with 30 points and 15 blocks with this strategy, but he almost might pick up a couple early fouls, and getting him on the bench is the only chance a big underdog has to hang with the Buckeyes.
Which is another reason why Texas A&M is a smart pick. Not that A&M would be a big underdog, but they’re an underdog against Ohio State for sure. Joseph Jones and Antanas Kavaliauskas draw fouls as well as anybody. Of course, they also pick them up in large numbers, too.
First round tidbit: Look for the Long Beach State/Tennessee game to be the fastest paced contest of the whole tourney.
Here’s the East Region:
2nd Rd Sweet16 Elite8 Final4 Finals Champs
1 N. Carolina 99.26% 81.43% 67.94% 48.69% 32.61% 20.90%
2 Georgetown 96.59% 81.33% 67.30% 32.43% 17.66% 9.09%
4 Texas 88.43% 55.85% 14.54% 6.11% 2.14% 0.69%
9 Mich. St. 60.71% 12.72% 7.24% 3.09% 1.10% 0.36%
3 Wash. St. 81.51% 52.09% 14.57% 3.27% 0.83% 0.19%
12 Arkansas 55.18% 24.33% 4.61% 1.52% 0.40% 0.09%
7 Boston Coll. 64.90% 13.48% 7.16% 1.44% 0.33% 0.07%
8 Marquette 39.29% 5.82% 2.70% 0.91% 0.24% 0.06%
6 Vanderbilt 71.57% 34.34% 7.68% 1.35% 0.27% 0.05%
5 USC 44.82% 17.69% 2.86% 0.83% 0.19% 0.04%
10 Texas Tech 35.10% 4.62% 1.79% 0.22% 0.03% 0.00%
13 New Mex. St. 11.57% 2.13% 0.11% 0.01% 0.00% 0.00%
11 GW 28.43% 8.04% 0.90% 0.08% 0.01% 0.00%
14 Oral Roberts 18.49% 5.52% 0.52% 0.04% 0.00% 0.00%
15 Belmont 3.41% 0.57% 0.09% 0.00% 0.00% 0.00%
16 E. Kentucky 0.74% 0.03% 0.00% 0.00% 0.00% 0.00%
This appears to be a two-team race but I think that Texas is undervalued in my system right now. UNC has quite the minefield to negotiate to get to the regional finals, with potentially scary games against Michigan State in the 2nd round and Texas in the 3rd round. Georgetown’s path is much easier with BC or Texas Tech and then likely Washington State or Vandy. However, the Tar Heels are still the most likely champion according to log5. Even so, there’s nearly an 80% chance they won’t be cutting down the nets.
First round tidbit: Arkansas was the most bizarre at-large selection, having a sub-.500 conference record in the embarrassingly weak SEC West. But the committee set themselves up to look good by pairing them with USC.
I’ll post some tourney-wide statistical miscellany tomorrow morning.
Tuesday, March 13, 2007
Bracket Breakdown: Midwest/West
Welcome to the kenpom.com Bracket Breakdown, a Boeheim-free zone. Log5 analysis is treated as gospel here, but I acknowledge that it’s not the answer to all of your bracket issues.
First up the Midwest Region:
2nd Rd Sweet16 Elite8 Final4 Finals Champs 1 Florida 99.72% 78.61% 56.54% 37.69% 21.07% 10.85% 2 Wisconsin 95.91% 68.27% 48.17% 25.15% 12.31% 5.45% 4 Maryland 86.42% 62.76% 26.04% 13.59% 5.64% 2.07% 6 Notre Dame 80.30% 47.69% 18.95% 6.86% 2.29% 0.66% 10 Georgia Tech 69.95% 25.14% 13.87% 5.07% 1.72% 0.50% 3 Oregon 86.02% 44.35% 15.76% 5.08% 1.51% 0.38% 5 Butler 72.39% 27.39% 7.13% 2.48% 0.65% 0.14% 9 Purdue 50.52% 10.89% 4.67% 1.70% 0.47% 0.11% 8 Arizona 49.48% 10.50% 4.45% 1.61% 0.44% 0.10% 7 UNLV 30.05% 6.16% 2.17% 0.45% 0.09% 0.01% 12 Old Dominion 27.61% 5.63% 0.70% 0.12% 0.02% 0.00% 11 Winthrop 19.70% 5.66% 0.84% 0.11% 0.01% 0.00% 13 Davidson 13.58% 4.21% 0.47% 0.07% 0.01% 0.00% 14 Miami (OH) 13.98% 2.31% 0.21% 0.02% 0.00% 0.00% 15 TAMU - CC 4.09% 0.43% 0.04% 0.00% 0.00% 0.00% 16 Jackson State 0.28% 0.00% 0.00% 0.00% 0.00% 0.00%
As we’ll come to see, this is the most wide open of the four regions, with six teams having at least a 5% chance to get to Atlanta, including an interesting oddball.
Florida’s path to a repeat is potentially quite a bit tougher than last season. Florida getting to the Final Four last season wasn’t a mammoth surprise. Last year at this time, I said:
Some would use Florida’s tournament history to disqualify them from a run to Indy. But don’t hold last season against them - they got the shaft by drawing a (then-)highly-underrated Villanova in the second round. With questions about the first and third picks in this region - the health of Villanova’s Allan Ray and the complete loss of shooting touch by Ohio State’s Je’Kel Foster - this is a good opportunity for Billy Donovan to restore his post-season image.
(Note: I actually said something like this about every team just to cover myself.)
Draw is so important, and the Gators were the second choice to get to the Final Four from their region according to log5. And once they got there, they had the good fortune to draw an 11-seed. A nice draw combined with great play is the formula for a national title.
The draw is not so bad this time around either, the 3- and 5-seeds are relatively weak and the 2-seed has a serious injury to deal with. You have to go to the 4-, 6-, and 10-seeds to find teams that could cause problems for the Gators. With that said, just based on numbers alone, they’re the least likely #1 seed to win it all.
Wisconsin has had major offensive problems since losing Brian Butch. But 2.93 of those 4.93 games have been against either Ohio State (10th best adjusted DE) or Illinois (3rd best AdjDE), so there’s reason for uncertainty about how much Brian Butch’s elbow has impacted the Badgers. Florida and Maryland have the best defenses in this region, and the Badgers would avoid them until the Elite 8.
Most folks know my feelings on Oregon. It started with the Maarty post, but for a while there have been signs that the defense is not top quality. The Ducks have a 1-in-20 shot of getting to the Final Four, and their defense has been better over the last four games, but one of those was against an Arizona team that is notorious for Pac-10 tournament flops.
Maryland gets a tough draw in the first round against Davidson. Davidson is 2nd in the nation in defensive rebounding and Maryland’s offense has been pretty dependent on offensive boards. If Davidson can take care of the ball against Maryland’s press, this game would get interesting.
While the selection committee has its principles, it seems to make arbitrary decisions more often lately. They gave more weight to what Butler did in November that February and March. But Butler has an explosive offense and was the victim of circumstance in its conference championship - a one-point loss on the road in which the opponent shot (eFG) 62.5% while attempting only four lay-ups. The Bulldogs had that luck returned by drawing ODU as a 12-seed, when they could have had to deal with Illinois or Arkansas. ODU is a poor three-point shooting team and they’ll be taking a lot of them against Butler.
As we get into the Cinderella seeds, two teams stand out - Notre Dame and Georgia Tech. Most folks would say it’s crazy to think one of them would make the Elite 8, and the odds are heavily against each doing so, but each is team is better than their seed would indicate. If they can get by serious first-round tests, watch out for these two. When you have Luke Harangody, anything is possible.
Now for the West:
2nd Rd Sweet16 Elite8 Final4 Finals Champs 1 Kansas 99.19% 78.52% 66.68% 48.48% 32.32% 18.89% 2 UCLA 98.08% 67.22% 38.23% 16.59% 8.01% 3.18% 6 Duke 86.21% 51.88% 27.54% 11.37% 5.22% 1.95% 3 Pittsburgh 89.71% 43.85% 20.50% 7.35% 2.93% 0.93% 7 Indiana 71.17% 26.83% 11.68% 3.76% 1.34% 0.37% 8 Kentucky 50.09% 10.76% 6.46% 2.69% 0.97% 0.28% 9 Villanova 49.91% 10.69% 6.41% 2.67% 0.96% 0.27% 12 Illinois 50.70% 30.03% 6.59% 2.29% 0.68% 0.16% 5 Va. Tech 49.30% 28.88% 6.21% 2.12% 0.62% 0.14% 4 S. Illinois 82.89% 38.49% 7.52% 2.36% 0.63% 0.13% 10 Gonzaga 28.83% 5.86% 1.43% 0.25% 0.05% 0.01% 11 VCU 13.79% 3.05% 0.49% 0.06% 0.01% 0.00% 13 Holy Cross 17.11% 2.60% 0.13% 0.01% 0.00% 0.00% 14 Wright St. 10.29% 1.22% 0.12% 0.01% 0.00% 0.00% 15 Weber St. 1.92% 0.09% 0.00% 0.00% 0.00% 0.00% 16 Play-In 0.81% 0.02% 0.00% 0.00% 0.00% 0.00%
The first thing that jumps out at me is how much easier Kansas’ side of the bracket is. The 2nd round game is no picnic, but the Sweet 16 game should be an easier test by going against one of [Virginia Tech, Illinois, Southern Illinois], although it that game should be very ugly in the points department. Putting that second game in perspective, the Jayhawks have the worst chance of all the No. 1’s to get to the Sweet 16, but the 2nd best to win the championship.
On the other side, while UCLA has a mini-home court advantage, they have to try to fend off some combination of [Indiana, Gonzaga] and [Pitt, Duke] to get to a potential game with Kansas.
The rep for Pitt is that they win with D, but they’ve given up more than a point per possession in eight of their last nine. Getting Aaron Gray over the 30 minute mark will be a key to their effectiveness.
The chances for miracle runs by Southern Illinois and Virginia Tech are severely stunted by having to face Kansas in the second round.
Duke and Indiana are given a real chance to advance deep. Another case of “I don’t see it”, but do you ever see a 6+ seed getting to the Final Four?
Kentucky and Villanova are looked at as underachieving power conference teams. Villanova has been playing better over the past month, while Kentucky’s defense has continued to regress. But Kentucky’s schedule has been the toughest in the land. The health of Mike Nardi could be crucial in deciding this one.
Finally, keep an eye on Illinois...only if you’re wistful for Dick Bennett’s days in the Big Ten. A great defense (3rd in adjusted DE) and a brutal pace (300th in adjusted tempo) have meant that the loser of an Illinois game has scored less than 50 in 7 of their last 13 games. You’ll be lucky to see either team get out of the 50’s in a dream Illini/Saluki 2nd round game.
Monday, March 12, 2007
Preview of a Preview
The Bracket Breakdown featuring log5 computations will commence tomorrow with the Midwest and West regions and continue Wednesday with the East and South. (Sneak preview: there are 10 teams with at least a 2% chance of winning it all if loyal reader Kevin Picklesimer’s calculations are correct.) I always pity the folks that are required to give instant reactions to the field because I need about 24 hours to process things.
Overall, I like what the committee did. My biggest beef is the overseeded (4)Virginia/(5)Tennessee/(12)Long Beach State/(13)Albany pod. I don’t get how one can defend seeding any of those teams that high, but one of them will get to the second week.
A couple of housekeeping notes: fellow rater Mike Greenfield is again running BracketBrains and if you want to test out how certain variables might influence tournament match-ups, then go sign up!
Also, for those that have been especially repulsed by my work this season - folks in the entire state of Nevada or the Stillwater metro area, for example - you’ve gotten your wish. I’ve been drafted into Dan Steinberg’s bracket contest and if things go well, I’ll lose to Jenn Sterger.
Saturday, March 10, 2007
Final Tally
Here’s the complete list of how teams with no rest fared the past two days. Overall, there was a slight underperformance.
Fatigued Team Rested Team Expected Outcome Villanova Georgetown L,61-65(33%) L,57-62 Syracuse Notre Dame L,78-80(42%) L,83-89 West Virginia Louisville L,64-66(38%) L,71-82 2OT Marquette Pitt L,64-67(36%) L,79-89 Cal UCLA L,59-71(10%) W,76-69 OT Washington Washington St. L,63-67(34%) L,64-74 Dayton Xavier L,64-75(15%) L,51-72 Fordham Rhode Island W,69-68(52%) L,69-73 St. Louis UMass L,62-69(23%) W,74-71 OT St. Joseph's George Wash. L,63-65(44%) L,48-58 E. Michgan Toledo L,62-72(16%) L,54-62 Ohio Miami L,60-61(44%) L,51-70 C. Michigan Akron L,60-75( 5%) L,53-82 W. Michigan Kent St. L,65-69(34%) L,66-75 Rice Central Fla. L,61-79( 5%) W,53-51 Southern Miss Houston L,68-70(43%) L,59-62 Marshall Memphis L,58-82( 2%) L,71-92 Tulsa Tulane W,69-66(64%) L,56-58 Kentucky Miss. St. W,74-71(60%) L,82-84 OT Arkansas Vanderbilt W,71-68(62%) W,72-71 Georgia Florida L,65-74(20%) L,57-74 LSU Ole Miss W,67-64(64%) L,60-80 Oklahoma Kansas L,58-68(15%) L,47-64 Texas Tech Kansas St. L,68-69(44%) L,45-66 Okla. St. Texas A&M L,61-74( 9%) W,57-56 Baylor Texas L,67-79(14%) L,69-74 Florida St. N. Carolina L,71-84(13%) L,58-73 Miami Boston College L,68-76(21%) L,71-74 OT NC State Virginia L,71-76(30%) W,79-71 Wake Forest Virginia Tech L,67-75(20%) L,52-71 Michigan Ohio State L,58-67(17%) L,62-72 Michigan St. Wisconsin L,57-59(37%) L,57-70 Illinois Indiana L,61-63(43%) W,58-54 OT Wins 10.1 7 Avg Score 65-71 62-71
Friday, March 09, 2007
No Rest, No Problem?
Several people, including myself, have noted that the log5 prediction formula doesn’t account for the situation of a fatigued team having to play a fresh team. It seems like a plausible idea, but looking at yesterday’s action, the jury’s still out.
The predictions below are based on yesterday’s ratings involving games where only one team had played the day before. And while only two wins were predicted for the fatigued bunch, we’d expect about five based on the sum of the win percentages. In aggregate, fatigued teams slightly underperformed, winning two or three fewer games than expected (although two of the fatigued winners were very unexpected). However, the average score wasn’t much different than expected.
I’ll update this for the 15 games that fall into this category today.
Fatigued Team Rested Team Pred Outcome Villanova Georgetown L,61-65(33%) L,57-62 Syracuse Notre Dame L,78-80(42%) L,83-89 West Virginia Louisville L,64-66(38%) L,71-82 2OT Marquette Pitt L,64-67(36%) L,79-89 Cal UCLA L,59-71(10%) W,76-69 OT Washington Washington St. L,63-67(34%) L,64-74 Dayton Xavier L,64-75(15%) L,51-72 Fordham Rhode Island W,69-68(52%) L,69-73 St. Louis UMass L,62-69(23%) W,74-71 OT St. Joseph's George Wash. L,63-65(44%) L,48-58 E. Michgan Toledo L,62-72(16%) L,54-62 Ohio Miami L,60-61(44%) L,51-70 C. Michigan Akron L,60-75( 5%) L,53-82 W. Michigan Kent St. L,65-69(34%) L,66-75 Rice Central Fla. L,61-79( 5%) W,53-51 Southern Miss Houston L,68-70(43%) L,59-62 Marshall Memphis L,58-82( 2%) L,71-92 Tulsa Tulane W,69-66(64%) L,56-58 Wins 5.4 3 Avg Score 64-70 63-71