Don’t Shoot the Prognosticator
"After the Crash" [Perspectives] places part of the blame for the current financial crisis on the software models created by the physics and math specialists whom Wall Street refers to as quants to police investment banks after the Securities and Exchange Commission (SEC) lifted a rule requiring them to maintain debt ceilings and federal reserves in 2004. Pointing fingers at the quants for their imperfect financial models is like blaming meteorologists for imperfect models of climate change or immunologists for their imperfect models of AIDS. Such scientists overwhelmingly understand that their models are imperfect and are constantly trying to better them.
The crisis was not caused by models run amok. The SEC decision the editors cite was made by politicians and bank management, none of whom was a practicing quant. The banks disregarded their risk management groups, and ambitious politicians and cozy relationships between lenders, servicers and government spurred on lending. Many quant models showed that the housing bubble was growing out of control as far back as 2005. Up and down the chain of leadership from Congress to the Bush administration to bank management to mortgage lenders and brokers to homebuyers the problem was ignored.
The editors acknowledge other causes of the crisis but fail to accept that the government owns the lion's share of the blame.
Gerald A. Hanweck, Jr.
New York City
Driving Out of Control
In "Driving toward Crashless Cars," Steven Ashley discusses next-generation automotive safety technology that takes various measures of control from the driver, including robotic cars capable of riding in close formation without any driver intervention. Apart from safety concerns, society has yet to utilize such technology to make the flow of traffic more fuel-efficient, a pressing need in this era. In particular, the infrared laser and microwave radar in speed-attenuating collision avoidance systems should be installed on all new cars to simultaneously slow down and space out vehicles. Networked-highway and GPS technology that reads digital signage and computes variable speed limits to slow vehicles ahead of freeway congestion is another fuel-saving and safety-enhancing innovation that could be deployed quickly.
Sherman Oaks, Calif.
Reading Ashley's article makes me want to keep my Saturn on the road for a long, long time. The idea of having to compete with a computer for control of a car sounds dangerous. At the very least, the "preset distance" to be maintained by a forward-collision warning system needs to be user-programmable. In Detroit rush-hour traffic, if you let your following distance open up to three quarters of a car length, someone will pull into it.
Hazel Park, Mich.
The "Earthrise" photograph taken of our planet from the moon during the Apollo 8 mission included in "Beacons in the Night," by John Rennie [From the Editor], reminds me of a question that has bothered me ever since I saw it in 1968: Why don't we see any stars in the picture?
Salt Lake City
RENNIE REPLIES: Many people have wondered about that discrepancy over the years. Stars aren't visible in the Apollo photographs because the surface of the moon and the earth itself appear so bright. Because of their brightness, the astronauts had to use short exposures. But the film could not handle the level of contrast involved in also picking up the images of the stars, which are many orders of magnitude less bright than the moon and the earth in the foreground both being illuminated by the relatively nearby sun. If the moon's surface were a dark color and the earth had not been in the sky, it might have been possible to photograph stars casually, but that was not the case.[break]
Try this experiment on your own: some clear night with a full moon, point a handheld camera at the sky and try to snap a photograph that simultaneously shows clear details of the moon as well as any stars in the field of view. Neither film nor digital cameras can routinely capture such disparate levels of brightness within a single frame.
In "Patternicity" [Skeptic], Michael Sher mer claims that humans evolved to "find meaningful patterns in meaningless noise," causing people to believe "weird things." As examples, he ascribes various irrational beliefs to "UFOlogists," "religionists" and "conspiracy theorists," thus exhibiting the very kind of false pattern recognition he claims to expose. Just as all snakes with red bands can be mistaken for the poisonous varieties, all religious people can be mistaken for those who see the Virgin Mary on the side of a building, all who seek evidence for extraterrestrial contact can be mistaken for those who see a face on Mars, and so on.
This hypocritical dynamic from the mouths of those who would circumscribe certain "types" of individuals who cannot be trusted to think is corrupting our public discourse about science and reason and injecting dangerously authoritarian modes of thinking into our political debate.
SHERMER REPLIES: Leith is right that it cannot be "authority" by itself that makes the distinction between a true and false pattern, between a false negative and false positive error, between meaningful and meaningless patterns, and so on. In politics, we decide on which pattern of government we want by voting. But in science, we rely on evidence, experimentation, corroboration, repeatability and the other tools designed specifically to avoid making those types of errors in pattern detection.
"Blocking Sound with Holes," by Charles Q. Choi [News Scan], gives incorrect dimensions for aluminum plates used in the described experiments. The plates were two to five millimeters thick and 20 centimeters wide.
"Turning Back the Cellular Clock," by Tim Hornyak [Insights],misstates experimental data regarding mice implanted with induced pluripotent stem cells (iPS cells): 37 mice received iPS cells made with the cancer gene c-Myc, and six died. Twenty-six received iPS cells made without c-Myc, and none died.