Tuesday, May 06, 2008

Last night in NBA playoff action, the Orlando Magic hit seven straight three-pointers in the third quarter of its game at Detroit. The Pistons ultimately won, however, 100-93. Jameer Nelson made three of the treys, Rashard Lewis made a pair, and Hedo Turkoglu and Maurice Evans each contributed one. All four of these players were roughly .400 shooters from behind the arc in the regular season, ranging from Nelson's .416 to Evans's .396.

The probability that, of the seven three-point attempts the Magic put up in the third quarter, all of them would go in, is calculated as (.400) raised to the 7th power. This roughly equals .002 or 1-in-500.

Overall, Orlando went 11-of-26 on three-pointers in the game. Using a more subtle approach, which takes into account the multiple opportunities for the Magic to have experienced a streak of seven straight made threes, yields a different probability for what Orlando did.

First, of the 26 three-point attempts, 19 were outside the scope of the streak of seven straight made ones. Of those 19, approximately 11 of them would be expected to be missed, given our assumed baserate of .400 success and .600 failure. Adding in the 1 from the formula linked above, we then multiple our previous probability estimate of .002 by 12, yielding .024. In other words, after each miss (as well as at the start of the game), there would have been a new opportunity to start a streak of seven made threes. Still rare, but not that rare (a conceptually similar analysis of mine from many years ago is available here).

No comments: