Good News for N.B.A. Fans

Tim Donaghy‘s 2007 arrest for betting on N.B.A. games, including games that he refereed, shocked basketball fans. Despite his astounding betting success rate (70 to 80 percent), Donaghy claimed that he never fixed N.B.A. games but rather used insider information, a claim that the N.B.A., the F.B.I., and the U.S. Attorney’s office were unable to disprove. Donaghy’s new book, Personal Foul, details his strategy, which he says relied heavily on personal biases among referees. If Donaghy is telling the truth about referee bias, the N.B.A.’s credibility is shot. But is he? TrueHoop blogger Henry Abbott investigates the stats and finds that Donaghy’s strategy doesn’t do a very good job of explaining his success. He writes, “Did Tim Donaghy fix games? His book is emphatic that he did not, but while it has provided many claims to that effect, it has not yet delivered evidence.” [%comments]

trader n

He probably did a bit of both. A diversified strategy is they way to go.


This is good news for the Thunderdom-ization of the NBA. Get rid of refs except where they carry a stretcher for those who get a certain foul. We get to repeat the exciting entertainment days of Rome while closing the gap on game fixing by the refs and the floppers.


The FBI reviewed videos of his games and they couldn't find proof that he ever fixed any games. There weren't any calls he made that they could point to. So he may actually be telling the truth about that part

Joe A.

There's other ways to bet than just betting on the winner with the spread. You can bet the over/under final score, etc...

Foul calls are very subjective in basketball. A game can be called 'tight', or the refs can 'let the players play'.

If a ref was betting on the over/under for the total score (Both team's scores added togther >200 points for example), it's very easy to affect the outcome without it seeming suspicious.
The ref, betting the under, lets the players play. Less foul shots, less stoppages of the clock, less total scoring.
The opposite stratgey for the over.

I think it would be very hard for an evaluator to determine whether or not this was being done- in the NBA the foul is like holding in the NFL and that old saw: There's holding on every play, it's just whether or not the ref wants to call it.

Having intimate knowledge of the other refs in the league and how likely they are to call a certain game tight or loose would be a huge advantage in this kind of betting.

Don't know if this is how Donaghy was betting though.


Jonathon K.

You don't need big questionable calls to turn the tide of a game. It's the NBA, there is travelling or a push foul on just about every play. A call on a certain player to put him in foul trouble and therefore on the bench could greatly change the game in one's favor.

It would be interesting to actually see his bets, and see where he took points, or would be the under, and see if trends develop there.


Regardless, it's good to know that federal dollars were used to investigate a game where grown men chase a ball. Nice.


That's because it's impossible to prove a negative.


Who cares? The NBA is the most boring league in the world because of the outsized impact dominant players have (i.e. Kobe & LeBron). 80 percent of the teams in the league entered the season knowing that they had pretty much zero chance at making the Finals, something which is not the case in baseball, hockey, or football.


You say "Despite [Donaghy's] astounding betting success rate (70 to 80 percent)" -- but actually in the article you linked to, that's just what Donaghy claimed was his success rate. And as the article goes on to argue convincingly, Donaghy's claims in general don't stand up to the evidence. Is there any evidence that that was ACTUALLY Donaghy's success rate?


If you want to actually see NBA fraud in action, just watch the 4th quarter of game 6 of the Western Conference Finals in 2002. It was blatantly fixed. The Lakers were actually injuring players without even getting called for a foul. The Kings were getting called for fouls against players they never even touched. End result? The Lakers had 27 free throws in the 4th quarter compared to the Kings 9, the series went the full 7 games, and the Finals was between two teams with vastly larger TV audiences than Sacramento.

That was the last time I paid any real attention to NBA basketball as a sport.


This question can be scientifically analysed. Get someone (or better, some-many) who is familiar with basketball and basketball refereeing, but doesn't know Mr Donaghy's betting record. Show them a whole lot segments from games Donaghy refereed (including ones he didn't bet on.) For each segment, have them pick what result they think the ref is aiming for on the assumption that there is a fix, or they can pick 'no fix' if it all seems fair. Compare the results to his betting record and see if there is a statistically significant correlation.

You'd need to control for false correlations due to people picking that the ref favoured whichever team got more points in the given segement, against Donaghy's alledged skill at picking winning teams (which are of course more likely to score more in a given segment.) For example you could show one person only segments in which one team scored 6 points and the other 4, and balance them so that there are an equal number of segments in which the eventual winning team is the 6 pointer, or the losing team. But then our knowledgeable viewer might remember those games and know the final winning team...

I'm pretty sure it could be done in a fair way, but it isn't simple.



Shouldn't this be something that can be proven using a scientifically based economic study. The data store should be vast, assuming it is publicly available. There are 30 NBA teams, playing an 82 game regular season schedule. That's roughly 1200 regular season games per season. I don't watch the NBA, but based on divisional groupings, some teams play others more often. If the ref-ing crews are different for each game, and the ref's don't always ref same team pairs (which I would think is a pretty strong certainty, then statistically significant information can possibly be divined from analysis of ref to player matchups in the data (versus other refs actions against a given player or the average of all refs against a player versus a particular ref).

So maybe we'll eventually get a refresh of the 1st freakonomics book with the sub-header that includes "What do the NBA and Sumo wrestlers have in common?"



His claims are nothing more than a diffusion ploy. Blame someone else so that they stop looking at you. Looks like everyone fell for an entertaining but unlikely reason for his high success rate.

Brings to mind two bastardized quotes

You can fool (substantially) all the prople all the time.


common sense ain't common at all