Math nerds

Texas blog Barking Carnival takes a look at the numbers and finds there is some relationship between switching to a fast-paced no huddle offense and a decline in your defensive efficiency numbers that year:

http://www.barkingcarnival.com/2016/10/26/13430628/trading-paces-can-swi...

(Note this isn't saying you can't have a good defense running HUNH, but that when you switch to it, you should expect a defensive decline that season)

A few months ago, there was some discussion about whether the "Hot Hand" was a real thing, or simply the expected result of chance over time.  A study in the 1980s ("The Cold Facts About the 'Hot Hand' in Basketball") suggested that believers in the hot hand were suffering from a "cogniitive illusion." 

In that mgoblog board discussion a few months ago, I mentioned that I had come across a study disputing that original study.  Unfortunately, I was unable to find it --  I had read it in a book that was given to me as a gift about a decade ago.  I had completely given up on finding the study.  However, I just moved into a new home last month, and while unpacking boxes this week, I came across the book!  It's titled "Anthology of Statistics in Sports", Edited by Albert, Bennett, and Cochran, and printed in 2005.  The specific study is titled "It's. Okay to Believe in the 'Hot Hand.'"  The authors' conclusion was that the original study was flawed, and that there was strong evidence that streak shooting was a real thing.

The data set included several games from the 1987-88 NBA season, and had several big name players included in the analysis.  One of those players, Vinnie Johnson of the Pistons, had a reputation as the ultimate streak shooter.  The authors looked to see if Vinnie really did accomplish low-probability streaks at higher frequency than other players, and the answer was a resounding "yes."  Fans were able to "make proper reputational attributions to those players who do the improbable and memorable more regularly than other players."

One of the more interesting results:  When looking at the probability to hit the next shot based on whether the previous shot(s) had been made or missed, Dennis Rodman's numbers really jumped out.  Probability after one make:  0.55.  After two makes:  0.78.  After 3 makes:  0.92.  Conclusion:  "success breeds success."  As he hits shots, his probability of a hit increases.  But then this:  Probability after one miss:  0.63.  After 2 misses:  0.69.   After 3 misses:  1.00.  Conclusioin:  For Rodman, "failure breeds success."  As he misses shots, his probability of a hit increases.  As with everything else concerning Rodman, that's just weird.  (Sample sizes diminished as the streaks continued, so this conclusion has to be taken with a grain of salt.)

Neat article about the "academic index", used by the Ivy League schools to ensure some kind of fairness across schools. The basic idea is simple: compute a formula based on GPA and SAT scores, and ensure each school has about the same average across their athletes.

More here:

http://www.nytimes.com/2011/12/25/sports/before-athletic-recruiting-in-t...

Should the Big Ten, SEC, etc., be forced to do something like this too? (it certainly would be interesting to know the AIs of various schools)