Nice idea BB but, sorry to say, it just wont work. Go to:
http://sportsdatabase.com/nba.py/query Enter: season=2009and team and (((P1+P2)<(o:P1+o:P2)) and (P3>o:P3)) into the SDQL box. This will tell you how many teams that year were losing at the half and how many times the losing team lost the 3rd quarter. In this example the losing team at the half lost the 3rd quarter 626 times. Then, conversely, enter: season=2009and team and (((P1+P2)<(o:P1+o:P2)) and (P3>o:P3)) into the SDQL box. This will tell you how many teams that year were losing at the half and how many times the losing team won the 3rd quarter. In this example the losing team at the half won the 3rd quarter 709 times. The overall winning percentage was 53%. With juice figured in it's a tough call with very limited profits if any. The Database offers stats from 1995 on. From 1995 through 2009 here is how 3rd quarter wins and losses stack up for the team that was losing at halftime. Wins: 9698 -- Losses: 8764 Winning percentage 525% Not too good.