Skip to main content

Posts

Showing posts from 2012

Poisson Games and Sudden-Death Overtime

Learning SQL

If you aren't aware of it, there's a free online course on databases (and SQL). I took it back when it ran live, but it's just as good self-paced. Jennifer Widom (Stanford) is an outstanding lecturer and the videos and assignments are excellent. SQLite is used to grade the online exercises, so I'd suggesting installing a local copy to experiment with as it's free. I'd also strongly recommend installing either MySQL or PostgreSQL (I recommend PostgreSQL; both are free) so you can learn while using a full-featured database server. BaseX is very helpful for learning XML and mastering XPath and XQuery for web scraping (also free).

https://www.coursera.org/course/db

http://www.postgresql.org/

http://basex.org/

BaseX has a module for handling JSON. I haven't used it personally, but it looks useful for learning about JSON.

http://docs.basex.org/wiki/JSON_Module

NBA Predictions for 11/21/2012

h_str = home team strength (including home court advantage)
o_str = opponent team strength (including away court disadvantage)
pr_home = estimated probability of home team winning

 home | opp | h_str | o_str | pr_home 
------+-----+-------+-------+---------
 ATL  | WAS |  1.03 |  0.93 |    0.85
 BOS  | SAS |  1.00 |  1.03 |    0.40
 CHA  | TOR |  0.99 |  0.96 |    0.63
 CLE  | PHI |  0.98 |  0.98 |    0.48
 DAL  | NYK |  1.02 |  1.08 |    0.28
 GSW  | BRK |  1.01 |  1.00 |    0.56
 HOU  | CHI |  1.02 |  0.97 |    0.70
 IND  | NOH |  1.01 |  0.95 |    0.72
 MIA  | MIL |  1.07 |  0.99 |    0.78
 MIN  | DEN |  1.01 |  0.99 |    0.58
 OKC  | LAC |  1.05 |  1.06 |    0.49
 ORL  | DET |  0.97 |  0.95 |    0.56
 PHO  | POR |  0.97 |  0.97 |    0.52
 SAC  | LAL |  0.95 |  1.01 |    0.28


Relative Strength of D1, D2 and D3 in NCAA Men's Basketball

Pooling inter-divisional results from 2002-2012, these were the relative strengths of the D1, D2 and D3 divisions for 2012.
Note that the rankings only use game results up to the end of February. I'm excluding results from March and April in order to test the quality of the subsequent predictions.
str = team strength ofs = offensive strength dfs = defensive strength sos = strength of schedule n = number of teams
team strength = (offensive strength)/(defensive strength)
In basketball, win probability is roughly
\[ \frac{{team}^{14}}{{team}^{14} + {opponent}^{14}} \]
where \(team\) is team strength and \(opponent\) is opponent strength.

 year |  str  |  ofs  |  dfs  |  sos  |  n    ------+-------+-------+-------+-------+------  2012 | 0.903 | 0.956 | 0.944 | 0.958 | 1021 (1 row)
 year | div |  str  |  ofs  |  dfs  |  sos  |  n   ------+-----+-------+-------+-------+-------+-----  2012 |   1 | 1.003 | 1.001 | 1.001 | 1.001 | 346  2012 |   2 | 0.902 | 0.960 | 0.939 | 0.958 | 281  2012 |   3 |…