Newmark's Door

Long but very worth-reading article on Three Mile Island and nuclear power in GQ. The heart of the piece:

At Three Mile Island, according to a 1980 inquiry by the Nuclear Regulatory Commission, the maximum level of radiation that anybody within a fifty-mile radius could have received from the accident was about 100 millirems—the equivalent of moving to Colorado for a year, or into a brick house for two. According to another study, by the Pennsylvania Departments of Health and Environmental Resources, among 721 locals tested, not a single one showed radiation exposure above normal. A similar study by the state’s Department of Agriculture found no significant trace of radiation in the local fish, water, or dairy products, which tend to register minute impurities. And a study released in 2000 by the Graduate School of Public Health at the University of Pittsburgh found that, twenty-one years after the accident, there was still no evidence of “any measurable impact” on public health.

Given the extreme scale of the meltdown at TMI—including an explosion of hydrogen, the liquefaction of radioactive uranium, and the release of a plume of radioactive gas into the air outside—it is reasonable to conclude that the lesson of Three Mile Island is not merely a matter of what went wrong at the plant but also an example of what went right. For so many people and so many systems to fail so spectacularly all at once, without any measurable effect on public health, may be the last, best proof that a system is working. . . .

When a meltdown like Three Mile Island scares us blind, creating an apocalyptic mythology about what happened there, we pay for our superstitions with sulfur fumes, global warming, and acid rain—our suicide pact with coal.

And when we fail to consider each of these issues with reason instead of fear, when we fail to make the tough comparison between nuclear power, with its potential for disaster, and coal plants, with their guarantee of it, this isn’t a reflection that we have no choices but that we refuse to make them.

It may be, more than anything else, an example of democracy working and failing at the same time.

"Casino insider tells (almost) all about security".

Even if you are cheating, and they know you're cheating, they might leave you alone if you're not that good at cheating. Take card counting: While counting cards in one's head is not illegal, a good card-counter in blackjack gains a statistical advantage over the house, and if the casino decides the counter is making too much money, he or she will be escorted off the premises.

But card counters have to be really good: One mistake an hour could swing the advantage back to the house. And casinos don't mind that. "If you're not perfect at card counting, you can still lose money," Jonas said. "They'll watch you count cards and if you make any mistake they'll just let you play."   

Charts, logos, and more.

Improve the world: teach a young person the fundamentals of personal finance.

(Link via Instapundit.)

Washington Post article claims that the supposedly horrible contest for the Democratic presidential nomination illustrates "the tragedy of the commons":

Individuals embroiled in similar dilemmas find them impossible to solve on their own, because they are confronted by a Hobson's Choice: Act selfishly and cause collective disaster, or act altruistically and aid someone else who is acting selfishly. Either way, selfishness wins.

"The way the system is set up, the more-selfish person has a higher probability of winning," social psychologist W. Keith Campbell said of the Democratic primary. "You end up with the more narcissistic, belligerent candidate."

Yes, but this problem is well understood and usually groups of people evolve norms or establish institutions to address the problem. It's only because the Democratic Party has arrogantly continued to tinker endlessly with their nominating rules in pursuit of "fairness" that it faces this problem today. 

Times staffers have to watch their Jacksons more carefully these days.

(Thanks to Michael Greenspan for the link.)

I don't imagine there's a good way to prove this, but the intuition seems sound: "More data usually beats better algorithms".

Dad takes seven-year-old son to a ballgame. Dad--he's a U. of Michigan professor who doesn't watch much TV--buys son "hard lemonade," not knowing that it has alcohol.

Policeman sends son to the hospital, and Dad (and his wife) lose custody of son for 48 hours.

It wasn't anybody's fault, you see. Everybody was just following "protocol".

J. K. Rowling sues to prevent a "reference guide" to Harry Potter from being published.

"Closely-watched case may spell trouble for software patents".

Most of these resources are pretty well-known; some of them have even been touted here. But it's nice to have links to them in a single place.

And here are Ngan Dinh's own suggestions. (You may remember her name because she has been mentioned on this blog before: here, for example, are her comments on Kevin Murphy's Applied Price Theory course at Chicago.) She is currently a visiting instructor at her alma mater, Bates College.)