Tuesday, January 22, 2008

Trend Data Trumps Big East Passion

Question: Should the Big East remain the 3rd-rated conference according to the RPI, what percentage of their teams will they advance? What does the trend data suggest?

YearConferenceFraction of ParticipantsPercentage (%)Seed Numbers
2007PAC-106/1060.02,3,3,5,8,11
2006ACC4/1233.31,3,4,10
2005Big XII6/1250.02,3,3,6,8,9
2004Big East6/1442.93,5,5,6,7,8
2003ACC4/944.42,3,6,9

Total Percentage: 26/57 (44.1%)

As evidenced by the trend data, the last two years have been statistical anomalies and deviate from the five-year average by more than 10 percent. Given this data, it would be reasonable to posit that the #3 conference (currently, The Big East) should gain approximately seven bids. Projections above eight bids or below six bids would be considered statistical outliers.

As for seeding, #11 is the lowest seeding to be gained over this five-year period (Stanford, 2007). Perhaps, the Big East seedings (3, 5, 6, 7, 9, 11, 13, 13) from the January 21 NCAA Tournament Projections may have been harsh. Nevertheless, the allocation percentage (50.0%) is more generous than the five-year average of 44.1%.

6 comments:

jameer1313 said...

No one argued with the number of Big East teams you put in your bracket. It was the ridiculously negative seeding due to your mid-major bias that people were upset about.

How the heck do you put Nova at an 11 seed? That's amazing, and I'm a Saint Joseph's (PA) fan.

Paymon said...

Jameer,

Here are the facts:
- 2 people had a problem with the "mid-major bias"
- The other person, Bryce, had 9 Big East teams in his latest bracket projection. The trend data disagrees with his assertions. Additionally, the Big East did not get 9 teams when they deserved it in 2006.
- As I admitted during this post, my seedings may have been harsh, especially for the back-end teams. They are revisited weekly.

The entire point of this post is to put aside passionate assertions (ahem, bias) and stick to the available data.

As more games are played, the seedings will normalize given that bracketologists will have more hard evidence from conference play.

My recommendation is that you visit Coolpohle's Bracketology and RPI Forecast. Villanova was in neither of their brackets according to the Bracket Matrix.

jameer1313 said...

I would love if Villanova missed the tournament. But that's not happening. This is a good basketball team with tons of talent. They are so much more than Scottie Reynolds. Pena and Cunningham in the front court (and if they ever get Drummond back). Those are big athletes. And Nova has more guards than Buckingham Palace.

Paymon said...

Touche. Talent means very little unless it gels into a cohesive unit.

Villanova has been everywhere on the performance stratum between great (15-min stretch against George Mason) and unwatchable (32 minutes of the LSU game).

jameer1313 said...

What do you think about Saint Joseph's (PA)? Any shot at the dance?

Evilmonkeycma said...

I was arguing from the point of view of the projection-prediction model. My bracket is not such a model, and I do not believe that 9 Big East will actually make the tourney. However, I still believe that those you had in were seeded too low, and I felt I gave an adequate defense of that assertion for Notre Dame, who I will not deny I am biased toward.

However, I felt the main problem was that several of the mid-majors were ranked too highly. A mid-major that completely dominates their conference will not receive the same boost from said domination that a major conference team would, simply because the competition is not as fierce (ignoring the A-10, which denies it is a mid-major anyway). In the case of Kent St. - I don't believe they would be a 8 seed even if they won out. However,it seems likely that they will not win out. That is why I had a problem with their ranking.