Lies, Damn Lies, and Statistics

• gorocks99

  

  I'm using kenpom.com final stats/rankings from the past 7 years (2004-2010), but here are some things to chew on:
  
  #1 ranked teams advance to the Elite 8 every time since 2004. Top 2 advance 86%, top 3, 86%, top 4, 79%, then 71%, 71%, 69%, 64%, 62%, 59% for top 5 through top 10 ranked. The top rated teams going into the tournament this year are:
  
  1. Ohio State
  2. Duke
  3. Kansas
  4. Texas
  5. Pitt
  6. SDSU
  7. Kentucky
  8. Purdue
  9. Wisconsin
  10. Notre Dame
  
  Only twice since 2004 has a team rated in the top 2 failed to advance to the Elite 8 – Kansas last year and Memphis in 2009. Only twice has a team rated in the top 5 failed to advance to the Sweet 16 – Kansas 2010 and Wisconsin 2008. That’s 94% likelihood that the top 5 teams will advance to the Sweet 16. Makes sense this year given the top 5 are top seeds plus Texas. But you’ve also got a 90% likelihood that the top 7 teams advance. Throw in SDSU and Kentucky and you’ve nearly got half your Sweet 16 done already.
  
  Potential teams to stay away from:
  
  9. Wisconsin: Although 9th or higher ranked teams comprise about 57% of the Sweet 16 pool since 2004, they are only 36% of the Elite 8. Wisconsin’s defensive efficiency (63rd) would rank them in about the 8th percentile of teams that have advanced to the Sweet 16, and in the 4th percentile of teams in the Elite 8. Not a likely candidate for either place.
  
  10. Notre Dame: 2 out of 7 teams since 2004 ranked 10th have made the Elite 8 – although half make it to the Sweet 16, and 30% of teams in the Elite 8 have been ranked 10 or higher, just one of those had an AdjD higher than Notre Dame’s is this season.
  
  19. Florida: 1 out of 7 teams ranked 19th have made the Sweet 16. Only 13% (7) of Elite 8 teams were ranked 19th or higher, and all but two of those had better AdjD than 2011 Florida.
  
  25. Arizona: Just 21% of Sweet 16 teams were ranked 25th or higher. Six of these (25% of that 21%, or around 5% of the total Sweet 16 teams) were worse defensively than the Cats.
• Prescott

  I think your work is very interesting, but not relevant. The difference between the top seeds and other teams is not as great as it once was. Some call it parity, I call it mediocrity.
• Mulva

  

  Somebody has to make the Elite 8 from the bottom half of the southeast. That region is impossible to predict.
• gorocks99

  

  Even "mediocre" teams in the past few years have been good in one way or another. Butler ranked 5th in defensive efficiency last year. Northern Iowa, the team that knocked off Kansas, was 16th. Cornell had a really good offensive efficiency and reached the Sweet 16 last year (they ranked 16th). While there's no magic formula to advancing in the tournament, you still have to be either stellar offensively, or stellar defensively, compared to your counterparts. Over 7 years the average AdjD ranking of a Final Four team is 10.4, and the average AdjO ranking is 13.4.
  
  If, for example, Notre Dame were to make the Final Four this year they would blow the curve wide open from where it has been. The lowest defensive ranking for a Final Four team since 2004 was 30th - that was MSU last year - and the median rank for those teams is 7.5. Notre Dame's ranking is 62nd heading into the tournament. That's a pretty strong outlier. Could they still make it? Absolutely, we're dealing with sports and strange what-the-eff-just-happened moments happen all the time. But from a statistical point of view it would be really, really odd if they did.
• vball10set

  my head is spinning--I'm just going to resort to the "throwing darts at the bracket" method
• thavoice

  I used to be a NCAA bracket nerd for a long time.
  
  I had taken like 15 years work of brackets and broke down teh records of each seed vs another seed. I cannot remember the exacts but there was an average of one per year of a lets say 11 beating a 6 or whatever...and I would select accordingly.
  
  It as some cool stuff! Kinda i gues.
  
  
  Then I grew up and got a life....but it was neat to say the leasaet.
• gorocks99

  

  I'm a market researcher, so predictive stats like this kinda are my life
• thavoice

  It def was interesting when I did all of that. The key was picking which of the upsets would happen. One year got lucky and got like 56 right out of the 63 I think games. One could say luck......and there was much of it...but there were some picks I did solely on the research...which I still think was pure luck. I think my research showed about a qtr of the time a 14 beat a 3 or something like that.....looked back 3 years and it not happend so I picked up taht year.
  
  Of course...it is just pure luck....but I said it was becuase of my research!
• swamisez

  

  I live by Ken Pom in the bracket pool. Won a couple last year. Would have won more if I wouldn't have chickened out and actually picked Duke.
  
  As much as I hate it, will be taking OSU this year due to Ken pom.