First sumo wrestling, now college basketball?

Here’s a link to a New York Times article discussing research by Justin Wolfers alleging point shaving in college basketball.

Justin Wolfers is also the co-author of an article that is highly critical of research purporting to show a big deterrent effect of the death penalty. I blogged about that issue a few days back.


spot

To me the most amazing stat in the paper was the fact that the favorites beat the spread in 50.01 percent of the games across the entire sample size. I really thought that overall the outcomes would just be more random than people expected and the underdogs would cover the spread more. Remember that the lines makers don't care about accurate lines- they only care about getting equal money on both sides so that they can make their money on the vig. I would also think that in most sports the larger the line is the less often the favorites would cover just from the simple standpoint of regression to the mean.

Robert Schwartz

I am not sure that the thesis is supported by the data, at least as far as discussed by the NYTimes today. There are lot of reasons why favorites don't cover large spreads that are not reflective of
illegality.

First, as you said in your article "It isn't clear why bettors prefer favorites, but such a tendency characterizes most betting against a spread." Bookmakers should therefore raise the spread on clear favorites in order to induce bettors to take the dog.

Second, if a game is out of hand the players and coaches on the winning side have no incentive to take risks in order to increase the margin of victory. Indeed, it can be time to rest the starters and play the scrubs.

Third, coaches who are playing teams that they will have to play again, have an incentive to keep the margin of victory down so as not to antagonize their opponents.

Read more...

abudai

Couldn't Wolfers have named names? What teams are the most egregious violators? Was the Fresno State team with Chris Herren and Dominick Young up there?

William Goodwin

I actually think there's little evidence to support the idea that "It isn't clear why bettors prefer favorites, but such a tendency characterizes most betting against a spread." In fact, as Wolfers' paper demonstrates, most betting markets are remarkably efficient. In the case of college basketball, it isn't simply that the spread evenly divides favorites and underdogs. It's also that the point spread serves as an "unbiased forecast" of the actual outcome of the games, meaning that bettors are collectively coming up with remarkably intelligent forecasts of the future.

PSD

Spot, it's interesting that you mention the 50/50 in terms of beating the spread.

I just finished James Surowieki's (? - no idea how to spell it) book, 'The Wisdom of Crowds.' In his book, James points out that sports betting (specifically football games) is empirically proven to be exactly a 50/50 split half the time the favourite will cover the spread, half the time it won't).

Surely, this random 'coin flip' nature is essential to getting equal money on both sides of the bet.

Anyway, sorry about this tangent, but yeah I definitely recommend this book for its insight into the collective wisdom of people vs. a few experts.

arespair

First off, the assertion that '...that the point spread serves as an “unbiased forecast” of the actual outcome of the games' is false.
Most sports books take a view and shade the line, particularly on big favorites where the public piles in - Duke, Kansas, Zags, UConn. Penn covered almost all their small Ivy spreads this year but almost none of the really big one - like the Harvard game cited. Are these teams that are large favorites, many of whom send multiple players to the NBA and other pro leagues shaving? According to the paper, there are 10-20 teams *per year* if you do the math that would have to be shaving as large favorites, every year, since 1989.

Secondly, using a 'web crawler' doesn't tell us if he used the opening, middle, or closing lines from covers.com. These spreads in large games move a lot, and 1-2 pt move to one side completely changes his analysis.

'Surely, this random ‘coin flip' nature is essential to getting equal money on both sides of the bet.'

This is also false. The books don't try to get equal money on both sides on all games, in fact, it'd be impossible to do so with up to 150 games a nite just on college hoops.

Going even further, if a sharp bettor came in to MGM and put $10k on Clemson +18 v Duke, and then 10 squares came in and put $10k each on Clemson, the book would be more likely to drop the spread, or leave it unchanged than move it up.

Who is making the bets is more important than how much.

For further reading, I recommend Chad Millman's excellent "The Odds: One Season..." for insight into this very topic.

freakonomics fan,
ares

ps Stephen, have you read Millman's book? Fascinating and a fast but detailed read.

Read more...

X-Tra Rant » Econometric Study Shows Point Shaving in College Basketball?

[...] Fascinating. Found via the Freakonomics Blog. [...]

Abnormal Returns » Bracketologists take note

[...] The Sports Economist also notes the NYT article and includes a link to the underlying paper. The Freakonomics guys, who know something about forensic economics, note the Wolfers research as well. [...]

Recurring Decimals….. » Economists on football

[...] Yes and in the NFL it sometimes takes a meeting of the United Nations Security Council to decide who had the final posession or if the ball is to be spotted on the 45.5 or 44.87654 yard line ! Besides if Sanderson is under the disillusion that ‘fixing’ does not take place in US sports - he needs to talk to his UofC colleague Steven Levitt. [...]

spot

To me the most amazing stat in the paper was the fact that the favorites beat the spread in 50.01 percent of the games across the entire sample size. I really thought that overall the outcomes would just be more random than people expected and the underdogs would cover the spread more. Remember that the lines makers don't care about accurate lines- they only care about getting equal money on both sides so that they can make their money on the vig. I would also think that in most sports the larger the line is the less often the favorites would cover just from the simple standpoint of regression to the mean.

Robert Schwartz

I am not sure that the thesis is supported by the data, at least as far as discussed by the NYTimes today. There are lot of reasons why favorites don't cover large spreads that are not reflective of
illegality.

First, as you said in your article "It isn't clear why bettors prefer favorites, but such a tendency characterizes most betting against a spread." Bookmakers should therefore raise the spread on clear favorites in order to induce bettors to take the dog.

Second, if a game is out of hand the players and coaches on the winning side have no incentive to take risks in order to increase the margin of victory. Indeed, it can be time to rest the starters and play the scrubs.

Third, coaches who are playing teams that they will have to play again, have an incentive to keep the margin of victory down so as not to antagonize their opponents.

Read more...

abudai

Couldn't Wolfers have named names? What teams are the most egregious violators? Was the Fresno State team with Chris Herren and Dominick Young up there?

William Goodwin

I actually think there's little evidence to support the idea that "It isn't clear why bettors prefer favorites, but such a tendency characterizes most betting against a spread." In fact, as Wolfers' paper demonstrates, most betting markets are remarkably efficient. In the case of college basketball, it isn't simply that the spread evenly divides favorites and underdogs. It's also that the point spread serves as an "unbiased forecast" of the actual outcome of the games, meaning that bettors are collectively coming up with remarkably intelligent forecasts of the future.

PSD

Spot, it's interesting that you mention the 50/50 in terms of beating the spread.

I just finished James Surowieki's (? - no idea how to spell it) book, 'The Wisdom of Crowds.' In his book, James points out that sports betting (specifically football games) is empirically proven to be exactly a 50/50 split half the time the favourite will cover the spread, half the time it won't).

Surely, this random 'coin flip' nature is essential to getting equal money on both sides of the bet.

Anyway, sorry about this tangent, but yeah I definitely recommend this book for its insight into the collective wisdom of people vs. a few experts.

arespair

First off, the assertion that '...that the point spread serves as an "unbiased forecast" of the actual outcome of the games' is false.
Most sports books take a view and shade the line, particularly on big favorites where the public piles in - Duke, Kansas, Zags, UConn. Penn covered almost all their small Ivy spreads this year but almost none of the really big one - like the Harvard game cited. Are these teams that are large favorites, many of whom send multiple players to the NBA and other pro leagues shaving? According to the paper, there are 10-20 teams *per year* if you do the math that would have to be shaving as large favorites, every year, since 1989.

Secondly, using a 'web crawler' doesn't tell us if he used the opening, middle, or closing lines from covers.com. These spreads in large games move a lot, and 1-2 pt move to one side completely changes his analysis.

'Surely, this random 'coin flip' nature is essential to getting equal money on both sides of the bet.'

This is also false. The books don't try to get equal money on both sides on all games, in fact, it'd be impossible to do so with up to 150 games a nite just on college hoops.

Going even further, if a sharp bettor came in to MGM and put $10k on Clemson +18 v Duke, and then 10 squares came in and put $10k each on Clemson, the book would be more likely to drop the spread, or leave it unchanged than move it up.

Who is making the bets is more important than how much.

For further reading, I recommend Chad Millman's excellent "The Odds: One Season..." for insight into this very topic.

freakonomics fan,
ares

ps Stephen, have you read Millman's book? Fascinating and a fast but detailed read.

Read more...

X-Tra Rant » Econometric Study Shows Point Shaving in College Basketball?

[...] Fascinating. Found via the Freakonomics Blog. [...]

Abnormal Returns » Bracketologists take note

[...] The Sports Economist also notes the NYT article and includes a link to the underlying paper. The Freakonomics guys, who know something about forensic economics, note the Wolfers research as well. [...]

Recurring Decimals….. » Economists on football

[...] Yes and in the NFL it sometimes takes a meeting of the United Nations Security Council to decide who had the final posession or if the ball is to be spotted on the 45.5 or 44.87654 yard line ! Besides if Sanderson is under the disillusion that ‘fixing’ does not take place in US sports - he needs to talk to his UofC colleague Steven Levitt. [...]