Monday, March 28, 2011

Tournament Summary - Final 4

Here are my estimates based on the matchup combinations we saw Sunday:

1 (Kansas) vs 11 (VCU) - 1 wins 90% of the time
2 (North Carolina) vs 4 (Kentucky) - 2 wins 60% of the time

After applying those numbers to Sunday's games, we get the following results:

Expected RecordActual Record% of Wins Expected
Colonial2.8 - 5.26-2214%
Pac 104.1 - 4.95-4122%
A102.7 - 3.33-3111%
SEC6.6 - 4.47-4106%
ACC7.7 - 4.38-4104%
Big Ten8.3 - 5.77-784%
Big East13.2 - 7.811-1083%
Mountain West4.8 - 2.24-383%
Big 127.0 - 3.05-571%
CUSA0.8 - 1.20-20%

Sunday, March 27, 2011

Tournament Summary - 6 Teams Left

Here are my estimates based on the matchup combinations we saw Saturday:

2 (Florida) vs 8 (Butler) - 2 wins 80% of the time
3 (UConn) vs 5 (Arizona) - 3 wins 60% of the time

After applying those numbers to Saturday's games, we get the following results:

Expected RecordActual Record% of Wins Expected
Colonial2.7 - 4.35-2185%
Pac 104.1 - 4.95-4122%
ACC7.1 - 3.98-3113%
A102.7 - 3.33-3111%
SEC6.2 - 3.86-497%
Big Ten8.3 - 5.77-784%
Big East13.2 - 7.811-1083%
Mountain West4.8 - 2.24-383%
Big 126.1 - 2.95-482%
CUSA0.8 - 1.20-20%

Saturday, March 26, 2011

Tournament Summary - Elite 8

And then there were 8.  It seems like just yesterday that we were starting this tournament.  Now, we are down to just 12% of the original field.  This group of 8 teams includes quite a bit of variety with 7 different conferences and 7 different seeds being represented.  Kansas is the lone 1 seed left, while Butler and VCU are showing that teams usually expected to fall the first weekend can put it all together and go on extended runs.  Here's a complete list of the remaining 8 teams:

1 - Kansas
2 - Florida
2 - North Carolina
3 - Connecticut
4 - Kentucky
5 - Arizona
8 - Butler
11 - Virginia Commonwealth

As for my conference breakdown, here are my estimates based on the matchup combinations we saw in this year's Sweet 16:

1 vs 12 - 1 wins 90% of the time
1 vs 5 - 1 wins 80% of the time
1 vs 4 - 1 wins 70% of the time
2 vs 11 - 2 wins 90% of the time
2 vs 3 - 2 wins 60% of the time
4 vs 8 - 4 wins 70% of the time
10 vs 11 - 10 wins 50% of the time

After applying those numbers to the Sweet 16 games, we get the following results:

Expected RecordActual Record% of Wins Expected
Colonial2.7 - 4.35-2185%
Pac 103.7 - 4.35-3135%
ACC7.1 - 3.98-3113%
SEC5.4 - 3.66-3111%
A102.7 - 3.33-3111%
Big Ten8.3 - 5.77-784%
Mountain West4.8 - 2.24-383%
Big 126.1 - 2.95-482%
Big East12.6 - 7.410-1079%
CUSA0.8 - 1.20-20%


Can we crown the Colonial as champs?  We sure can!  With VCU matching up against KU in the Elite 8, even a loss would only drop the CAA down to 179%.  Unless Arizona plans on winning the national championship, this is going to be a runaway victory for the CAA.

On the other end of the spectrum, we have a 4 way race (not counting CUSA).  Both the Big Ten and Mountain West have run out of teams, so their numbers are fixed.  With KU still representing the Big 12 and Connecticut still representing the Big East, both of those leagues have a chance to move up and shed the label "disappointment".

The biggest mover in the last couple of days was the SEC.  Kentucky's upset of Ohio State, along with Florida continuing to take care of business, has led the SEC up in the rankings.  No longer appearing to be a disappointment, the SEC now has a chance to make a case as one of the more impressive performances by a conference.

As always, stay tuned for more analysis as the field continues to shrink...

Friday, March 25, 2011

Conference Update - 12 Teams Remaining

This will be quick.  Expect more following tonight's games as we prepare for the Elite 8.

Updated Conference Standings (% of expected wins):
Colonial - 182%
Pac 10 - 132%
ACC - 125%
A10 - 115%
Big 10 - 92%
SEC - 89%
Mountain West - 83%
Big East - 80%
Big 12 - 77%

Teams Remaining:
ACC - 2
Big East - 2
SEC - 2
A10 - 1
Big 10 - 1
Big 12 - 1
Colonial - 1
Horizon - 1
Pac 10 - 1

Monday, March 21, 2011

Tournament Summary - Sweet 16

One weekend down and we are now left with 16 teams standing.  Surprises abound as 12 seed Richmond, 11 seeds Marquette and Virginia Commonwealth, 10 seed Florida State, and 8 seed Butler all advanced to the 2nd weekend.  We also saw #1 seed Duke and #2 seeds San Diego State and North Carolina win by the narrowest of margins.  Interestingly enough, the majority of these upsets have all occurred in the same region, the Southwest.  #1 seed KU currently has a path to the Final Four that would include seeds 16, 9, 12, and then 10/11, making them a near shoe-in to be visiting Houston this year.  The West region came the closest to following expectations as #5 seed Arizona is the region's biggest Sweet 16 "surprise".

In order to provide an update to my conference breakdown I needed to determine the likelihood of each team advancing.  Unlike the first round where a 1 always plays a 16, 2nd round matchups can be different combinations of seeds.  This makes it much harder to use past data to predict the odds of any given team advancing.  My solution was to attempt to establish percentages on my own.  You may agree or disagree, but I think these should be close enough to accomplish the desired goal.  Anyway, here are my estimates based on the matchup combinations in this year's round of 32:

1 vs 8/9 - 1 wins 80% of the time
2 vs 10 - 2 wins 80% of the time
2 vs 7 - 2 wins 70% of the time
3 vs 11 - 3 wins 80% of the time
3 vs 6 - 3 wins 60% of the time
4 vs 5 - 4 wins 50% of the time
12 vs 13 - 12 wins 60% of the time

After applying those numbers to this weekend's games, we get the following results:
 
Expected RecordActual Record% of Wins Expected
Colonial2.2 - 3.84-2182%
ACC4.9 - 3.17-1143%
A102.6 - 2.43-2115%
Pac 103.5 - 3.54-3114%
Mountain West3.8 - 1.24-1105%
Big Ten6.9 - 5.17-5101%
SEC4.5 - 2.54-389%
Big 125.2 - 2.84-477%
Big East12.1 - 5.99-974%
CUSA0.8 - 1.20-20%


The first thing that jumps out at me is the ACC.  Their 7-1 record looks amazing until you realize that they were basically expected to be 5-3 at this point.  Yes, they are doing much better than expected, but just because everyone on TV is hyping their start through the first couple of rounds, don't crown them as the winners of the "most overachieving performance by a conference" just yet.  In reality, the Colonial has the lead in that race following VCU's pounding of Purdue to reach the Sweet 16.

As for disappointments, a lot has been said about the Big East and their 9-9 record.  Although they are indeed well below their expected win total of 12 games through the first weekend, they aren't the only conference to disappoint so far.  The Big 12 currently sits ever so slightly above the Big East in my standings, while the SEC is only a disappointing finish away from possibly taking the "honor" of most disappointing.  (I'm leaving out CUSA because of their small sample size and lack of quality seeding.)

Stay tuned for more analysis as the week rolls on...

Saturday, March 19, 2011

Tournament Summary - Down to 32 teams

We are now down to 32 teams still fighting to win and advance further in the tournament.  After one of the more competitive Thursdays of all time, Friday wasn't all that bad either.  Between the two days combined, 10 of the 32 games were decided by 4 points or less.  Technically, we saw 7 upsets, although most would agree that the 7/10 and 8/9 games are virtual tossups.  Outside of those seeds, 11 was apparently the seed to be as 3 of the 4 made it to the round of 32.  A 10 seed, 12 seed, and 13 seed also survived to play during the weekend.

As you watch games today, you will likely see a breakdown of records by conference.  With the Big East leading the way with 11 teams, they will likely get a lot of the focus.  They posted a respectable 7-4 up to this point, but are they the most successful conference right now?  How do we determine that?

For example, the Big Ten's record looks pretty solid at 5-2 so far.  However, with 3 protected seeds amongst the group, you would expect to see a handful of early wins.  In order to truly see how well a conference has done so far, you must compare them to some sort of baseline expected result.

If you look back at the seed-by-seed results ever since the field expanded to 64, you see the following 1st round winning percentages:

1 seeds - 100%
2 seeds - 96%
3 seeds - 85%
4 seeds - 79%
5 seeds - 67%
6 seeds - 69%
7 seeds - 63%
8 seeds - 46%
9 seeds - 54%
10 seeds - 37%
11 seeds - 31%
12 seeds - 33%
13 seeds - 21%
14 seeds - 15%
15 seeds - 4%
16 seeds - 0%

Based on those percentages, a conference like the A-10 (with seeds of 6, 7, & 12) would be expected to win 1.7 games.  This is easy to calculate.  6 seed = 0.69, 7 seed = 0.63, 12 seed = 0.33

.69 + .63 + .33 = 1.7 wins

In reality, the A10 has gone 2-1, so they have slightly outperformed their expected 1.7-1.3 record.  By running the numbers on all the conferences, we can see which conferences have truly over performed and which ones have been disappointing.

Expected RecordActual Record% of Wins Expected
Colonial1.8 - 2.23-1167%
ACC3.2 - 1.84-1125%
Pac 102.4 - 1.63-1125%
A101.7 - 1.32-1118%
Big Ten4.4 - 2.65 -2114%
Big East8.0 - 3.07-488%
Big 123.4 - 1.63-288%
Mountain West2.3 - 0.72-187%
SEC3.3 - 1.72-361%
CUSA0.8 - 1.20-20%


A few notes on the table above:
  • % of Expected Wins is calculated as (actual wins) / (expected wins)
  • First Four games were included.  In those cases, games were between evenly matched seeds, so I simply gave each team a 50% chance of winning.  This is why the Colonial has played 4 games with only 3 teams.  (VCU won the First Four game against USC and then beat Georgetown in the round of 64)
  • Only multibid leagues were considered.  One bid leagues are way too small a sample size.  For example, Morehead State won as a 13 seed, giving the OVC a 303% in the far right column.

Thursday, March 17, 2011

It's Tournament Time!

Selection Committee, you're history!  Well, atleast until next year.  The seeds have been set, and regardless of what we think of them, it's time for some games to be played!  In fact, the madness has already started.

As the tournament progresses, March_24_7 will be digging through the results to bring you interesting trends and analysis that you won't see anywhere else.  Look for another update following the conclusion of the first round games.

Wednesday, March 16, 2011

A Respectful but Honest Letter to the Selection Committee

Dear Selection Committee,

I greatly appreciate all the hard work you just went through in order to select the field of 68 for the 2011 NCAA Men's Basketball Tournament.  I know that you spent hours studying numerous resumes, probably to the point where they all began to run together.  I believe your intentions were honorable and your work ethic commendable, but I would like to make some observations on the results of your labor.

1) Your top 4 seeds were rather good.  Arguments could be made to make some minor changes, (surely Purdue's 2 top 25 wins and poor finish should not have placed them ahead of Louisville and their 6 top 25 wins) but I wouldn't move any of your top 16 teams up or down by more than 1 seed line.  Considering that you sometimes have to deviate from your S-Curve in order to follow your bracketing rules, I rate your job in this area as an "A-".

2) Giving USC a bid and leaving out Colorado is difficult to understand.  Both had 5 top 50 wins.  Both went 3-3 against teams in the 51-100 range.  Both made it to the semifinals of their conference tournaments.  Colorado had 3 losses outside the top 100 while USC had 6, including 3 losses to teams outside the top 200!  What exactly did USC do better than Colorado to earn a bid?  Maybe it was the Trojans signature win over Texas?  Nope, Colorado beat that same Texas team too.  In the end, this selection is rather puzzling, and I am giving you a "D-".

3) Your favoritism to the Big 10 was not all that subtle.  Penn State, with 14 losses and a losing record vs the top 100, as a 10 seed?  Illinois, who finished 6-10 in their last 16 games and only won 3 of their final 11 games against tournament teams, as a 9 seed?  Finally, how did you conclude that Michigan deserved an 8 seed?  Was it their 0-8 record vs teams that you seeded in the top 8 of the tournament?  Was it their additional 5 losses to teams not seeded in the top 8?  You know that makes 13 losses overall?!  For seeding the Big 10 excessively high, I give you a "D+".

4) Although you failed with the Big 10, I would like to congratulate you on your Big East selections.  First of all, thank you for setting a record by selecting 11 teams from one conference.  Although many people were whining and moaning about this beforehand, it was the right move, as all 11 were clearly tournament worthy.  I would also like to congratulate you on your seeding of these 11 teams.  Pittsburgh was a no brainer for #1 seed.  Notre Dame, UConn, and Syracuse should all feel like they were fairly treated by being handed 2 & 3 seeds.  Louisville may have wanted a higher seed following their run to the Big East Tournament Finals, but they were given a protected seed and should have no major complaints.  St. John's, West Virginia, Cincinnati, and Georgetown were all rewarded for having good seasons.  They will have every chance to succeed from 5 and 6 seeds.  Marquette may be a little underrated as an 11, and I would have probably flipped them with Villanova's 9 since the Wildcats finished so poorly, but each of these teams is in the tournament despite quite a few negatives on their resumes, so I don't see too much room for complaining.  Overall, I give you an "A" on your seeding of the Big East.

5) Unfortunately, my next point is once again negative.  You gave Utah State a 12 seed.  Did you realize that they went 30-3 and survived with only 1 loss in a reasonably competitive conference?  Did you notice that those 3 losses included a road loss @ 3 seed BYU and a road loss @ 6 seed Georgetown?  Those are places that almost no team in the country would be favored to win!  Utah State had one loss, once again on the road, to a mediocre Idaho team, but they also had 30 wins to help counter these losses.  You might say that tournament teams should dominate teams in the 100+ range of the RPI and that Utah State was only taking care of business.  Well, in a sense you are right, but then explain to me how Utah State, who went 27-1 against teams ranked 100+, ended up a full seed line behind 11 seed USC.  Did you forget that the Trojans were only 11-6 vs these same 100+ ranked teams?  If I am piling on, I'm sorry, but truth be told, you screwed up big time on this one.  Did you know that there was a collection of 89 bracketologists whose projections were compiled and studied?  Did you know that there wasn't a single projection of Utah State being worse than an 11?  Did you know that 89% of these bracketologists had Utah State's seed in the single digits?  How do you defend yourself being such an extreme outlier?  Personally, I can't defend you, and on this one, I give you an "F".

6) Very few people pay much attention to the 13-16 seeds.  These are automatic qualifiers, so everyone already knows they are in the field and usually just assumes that it doesn't matter where they are seeded.  They will probably lose in the first round.  Well, I believe your job is to correctly seed all 16 lines, and I have an issue with your choice of Morehead State over Bucknell for a 13 seed.  You do realize that Bucknell won @ 12 seed Richmond, right?  Morehead State's best win was @ 14 seed Indiana State.  Advantage Bucknell.  Maybe you liked Morehead State's 12-4 record vs teams ranked 200+.  If so, wouldn't Bucknell's 18-2 record against those same teams be even better?  Oh, I know, the real difference is in the 100-200 range competition.  Let's see... Morehead State went 7-3 and Bucknell went 6-1.  Nope, once again advantage Bucknell.  The only other possible advantage Morehead State might have is that they played a good close game @ Florida before eventually losing.  It is true that Bucknell didn't come close to pulling any big time upsets, but answer this question for me... Why does a close loss to a quality team help Bucknell but not Utah State?  I thought we already established that close losses are still losses.  What gives?  I'm going to have to give you another "D-" on this one.

7)  Here's another head to head comparison.  Team A lost 1 game outside the top 100.  Team B lost 2 such games.  Team A was 8-1 vs teams 51-100.  Team B was 4-2 against those same teams.  Team A had 6 top 50 wins.  Team B had 3.  Team's A & B both play in the same conference.  Team A took 2nd place in the regular season and then won the conference tournament.  Team B took 1st place in the regular season, but then lost in the semifinals of the conference tournament.  In case you didn't know, Team A is Old Dominion, and you seeded them as a 9.  Team B is George Mason, and you seeded them as an 8.  Is winning a regular season championship enough to trump nearly every other category combined?  I don't think so, and for this, I give you a "C-".

8)  You say that you don't take into consideration past performances when making your selections.  If that is true, then why did Butler receive an 8 seed?  They were 1-3 against tournament teams and they had 6 losses to teams that didn't make the field of 68.  Did you know that they lost to a Youngstown State team that only won 6 other games all year?!  Did you realize that you seeded Butler ahead of Old Dominion, despite ODU winning more games against a tougher strength of schedule?  How do you explain that one?  Did you fall in love with Butler's 9 game winning streak to finish the season.  If so, you must have forgotten that ODU also won their final 9 games, and ODU played better competition in the process.  Oh, one more thing.  Remember that stat about Butler winning only 1 game against tournament teams?  Well Old Dominion won 6 such games!  This choice really makes me question your focus on this year only, and for that, I give you an F.

9)  Those of us who like to call ourselves "bracketologists" look forward to Selection Sunday as a chance to validate our hard work.  We enjoy the thrill of trying to get inside your heads and predict what you will do.  In fact, a website called The Bracket Project collected the results of 89 different bracketologists and compared their predictions to the results of your work.  Do you know how many of these 89 selected the same field of 68 as you did?  Do you know how many selected Utah State as a 12 seed?  Do you know how many selected Michigan as an 8 seed?  The answer to all these questions is NONE!  Not even one!  How can you possibly excuse yourself for being such an extreme outlier?  For goodness sake, there was a bracket that gave Syracuse a 1 seed.  Another bracket didn't even give UCLA a bid!  These weren't a group of 89 mindless robots spewing out identical information.  This was 89 individuals throwing out brackets that were all over the spectrum, yet somehow you managed to produce seedings that none of the could.  How can you possibly defend yourselves at this point?  For these 3 examples, you deserve an "F-".

Some say that we need more "basketball minds" on the Committee.  Others say that the tournament should start off including everyone so that we don't even need a Committee at all.  Personally, I don't like either of these ideas, but I can't say that I like a lot of the current status quo either.  For next year, will you please make a pledge to all of us basketball junkies?  Will you pledge to use common sense?  Will you pledge to look at a team's entire resume, and not just one area?  (Utah State's lack of quality wins, USC's abundance of quality wins)  Will you pledge to perform sanity checks to compare teams head to head in order to make sure your seeds make sense?  (Old Dominion vs George Mason, there is no reason that George Mason should be seeded higher)

As a bracketologist, I look forward to being as accurate as possible.  Unfortunately, that partially depends on your ability to produce a quality bracket.  I stand by my bracket and feel that I can defend the seeding of every team in my field of 68.  Can you say the same?

For all bracketologists, as well as college basketball fans everywhere,

March_24_7