The Wisdom of Crowds IV (Superbowl LXII)

February 5, 2008
Posted by Jay Livingston

1. The Statistical Wisdom of Oddsmakers.

Andrew Gelman writes:

if you look up "football" in the index of Bayesian Data Analysis, you'll see that football point spreads are accurate to within a standard deviation of 14 points, with the discrepancy being approximately normally distributed. So, a 14-point underdog has something like a 15% chance of winning. It's funny how people don't get this sort of thing.

The Giants were a 12-point underdog. The money line was about 9:2 – very close to the line Andrew would have set given the point-spread. Do bookies all have well-thumbed copies of Bayesian Data Analysis on their bookshelves?

2. The NonWisdom of Oddsmakers, the Wisdom of Crowds.

The oddsmakers set the opening line at 13 ½, but not because they thought that represented the strengths of the teams. They thought that the “true” line should be 12 or 12 ½. They reasoned that a lot of unwise bettors – people who bet on only the Superbowl and don’t know much about football – would be bet New England. The Patriots after all were undefeated, the Team of the Century, and all the rest of the hype. These bettors, so the logic went, would bet the Pats even at the inflated line.

But from the start, the supposedly naive money came in on the Giants. The line came down, and people still bet the Giants. The bookies took a bath. It happens.

3. Local Color.

David Tyree, the Giant who made The Catch, is a graduate of Montclair High School.

For other posts in this blog on football, betting, and the wisdom of crowds, go here.

Cultural Literacy

February 2, 2008
Posted by Jay Livingston

College students know about Lolita, at least at the high-SAT campuses. Apparently the same cannot be said for the folks at Woolworth's in the UK. Until the flak hit the fan, they had been offering a bed-desk-cupboard unit for girls age six or so. Nothing wrong with that. Except the name of the unit was the Lolita Midsleeper Combi.


British mums in an online chatroom didn’t think it was such a good idea to give girls’ bedroom furniture the name of a sexually precocious fictional twelve-year-old. The Internet makes it much easier to organize this kind of protest, and Woolworth’s discontinued the item.
But how had it slipped through in the first place? According to the Times,

"What seems to have happened is the staff who run the Web site had never heard of Lolita, and to be honest no one else here had either," a spokesman told newspapers.
"We had to look it up on (online encyclopaedia) Wikipedia. But we certainly know who she is now."

The Sun (“Fury at Woolies Lolita Girls Bed”) added

Woolies confirmed the bed had now been axed. A spokesman said: “We will be talking to the supplier with regard to how the branding came about.”

I’ll have to ask Claude the brand consultant about this.

Meanwhile, it’s not the first time pedo-ignorance has caused embarrassment in the UK. In its Christmas marketing, the website for Tesco, a large British retailer, had a Toys and Games section which offered Legos and Barbies and other stuff you’d expect. It also included a stripper pole, with the message, “Unleash the sex kitten inside ... soon you'll be flaunting it to the world and earning a fortune in Peekaboo Dance Dollars.”

After complaints, they moved the item to the “fitness” section.

Seeing Through the Clouds

January 31, 2008
Posted by Jay Livingston

Tag clouds offer content analysis at a glance. Here, for example, is the cloud of Monday’s State of the Union speech.

The biggest tags are no surprise: America, Iraq, people, terrorists. Nor are hope and future, which fit with the often-noted American cultural traits of optimism and future orientation (besides, it might have been difficult for Bush to look backward and review the list of his accomplishments).

Two tags of about the same size as these got my interest: trust and world. Here, you have to look at the contexts, and you have to look at what does not appear in those contexts, to understand what they mean in the Bush perspective.

Trust is usually a reciprocal sentiment, and Bush might have stressed the trust that people must have in their government, particularly in time of war. Or he might have said something about his administration having kept the trust of the American people (or would that have been too much of a stretch even for Bush?).

Instead, the speech was all about the government trusting the people. In fact, lurking not very deep in the subtext is the idea that the government itself is not to be trusted, certainly to be trusted with money (all those earmarks). This trust-the-people theme is, of course, just a flattering way of saying that the government is not going to do much for the people. Instead, the Bush administration will trust us to take care of ourselves as best we can.

The Bush view of “the world” is similarly unreconstructed. American exceptionalism reigns. No “taking our place among the nations of the world” or “working together for a better world.” Instead, Bush looks at America and the world as though he were a fan at a football game: We’re number one, and they’re out to get us. Cooperation with other countries is not an option, unless, of course, they want to co-operate by doing what we tell them.

“We showed the world the power and resilience of American self-government.” Take that, world.

I haven’t seen any polls yet measuring popular response to the speech. But I wonder if this stuff still sells.

When the Average Isn't Average

January 29, 2008
Posted by Jay Livingston

Here’s George Bush in the State of the Union last night urging Congress to make his tax cuts permanent.

Unless the Congress acts, most of the tax relief we have delivered over the past 7 years will be taken away. Some in Washington argue that letting tax relief expire is not a tax increase. Try explaining that to 116 million American taxpayers who would see their taxes rise by an average of $1,800.

You can’t blame a guy for trying, and you can’t blame the general public for not appreciating why the “average” tax cut is not the same as the tax cut for the average person. But income and income tax distributions are so skewed that using the average is misleading. When Bill Gates walks into the room, the average income goes way up, but everyone in the room except Gates is now below average.

The Tax Policy Center has computed the tax savings for each income quintile. For the middle income quintile, those tax cuts meant a savings of, on average, $814. In other words, the median savings was less than half the mean that Bush mentioned.

The big winners in the tax cuts are those at the top. The savings for the top fifth averaged $7452. For the top 0.5%, the tax cuts have meant an annual savings of over $100,000.