Sunday, January 3, 2010

Some suggested research for the NBA

As a quick follow-up to my earlier post re: the Pedowitz Report, here are three things the NBA could entertain if they wanted substantive evidence of whether former referee Tim Donaghy altered game outcomes to advance his betting propositions.

1. My first suggestion, ironically, was inspired by the NBA's own inquiry into the betting scandal.  In the Pedowitz Report (pp.113-14, under "VII. Recommendations; 4. Gambling Enforcement, Detection and Deterrence; c) Gambling Monitors; and d) Statistical Screening for Gambling and Bias"), the following is respectively discussed (emphasis added):
The League has now arranged to obtain information on a regular basis from individuals and entities involved in the gambling business who can provide the League with information about unusual movements in the betting lines, rumors about things such as injury reports or referee schedules or where the “smart money” is being wagered. By flagging games or individuals for the League to investigate, these monitors may help the League detect gambling or misuse of confidential information....
Since the 2003-2004 season, the League has been collecting data on calls and non-calls for each referee. The collection system was designed by Sibson as part of the overall effort to redesign the officiating performance program. The system itself was built by the League. Although this system was developed for training and instructional purposes, we have worked with the League and Sibson to develop a prototype, proprietary system for screening games in an effort to help detect data patterns that may suggest misconduct by referees and others. Data ― including this foul call information and the movements of betting lines ― can be analyzed using various algorithms to flag patterns consistent with questionable behavior. While this system is in development, the League has already started to actively monitor several high level data-points (such as line movements) for every game for signs of potential misconduct, and certain game and betting information is distributed to League management on a daily basis. For those games that are flagged, the League has undertaken further review. In addition, the League hired Steven Angel, a former consultant with Sibson, as Senior Vice President for League Operations and Officiating to, among other things, help coordinate wagering intelligence and game screening...
This collection of gambling and referee behavior data seems prudent, as do the related assessments.  Why, though, restrict such analysis to current and future activity?  If the referee/call data has been collected since the 2003-04 season (which, interestingly, is when former referee Tim Donaghy claims he first bet on NBA games), why not perform these analyses on Donaghy's games (and others) beginning with the '03-'04 season?

An interesting footnote related to this suggestion: According to Ken White, CEO and lead oddsmaker for Las Vegas Sports Consultants, the world's largest oddsmaking company, he researched betting trends involving Tim Donaghy and submitted his report to the NBA in the fall of 2007.  Of this situation, White told the Las Vegas Review-Journal, "They never called back to discuss it or anything."

2. I have written previously about Indiana University Business Professor Wayne L. Winston's commentary on whether Tim Donaghy fixed games.  Additionally, in his book, Mathletics, he suggests (p.247) we should:
compare the percentage of fouls Tim Donaghy called in games where the Total Line increased by at least two points to the percentage of fouls he called in all other games.*
I would add this analysis should be conducted dating back at least to the 2003-04 season.  Serious NBA scandal observers are likely noting to themselves that Donaghy claims to have never bet on totals ("over" or "under" total points in a game), and thus may be discounting Dr. Winston's suggestion.  I think Winston's proposed analysis would, in fact, be telling, for reasons I will explain in due time.

*A move of two points or more in the Total Line is generally considered a big, or unusual, move.

3. The third thing the league could do is either the most simple or the most difficult: identify and locate big-time, heavy-hitting professional gamblers, and then interview them about the NBA betting scandal.  Some of these individuals have done their own - remarkably sophisticated - assessments, and there may be considerable wisdom to gain from speaking with them.