Blame the modeler, not the model
In your editorial about models developed by mathematicians and statisticians in the financial industries (“When you can’t believe the model,” Feb. 19), you point out several models you say failed. My belief is that many of those individuals and their models predicted the present situation, but those predictions were ignored. In addition, I would also bet there were many instances when it was necessary to build models that met the needs of upper managers to show unlimited growth. Unfortunately, when models predict an outcome that requires caution and restraint, they can be easily dismissed or modified. If our current financial mess was a matter of safety and people had been physically harmed, you can believe there would be an 80-mile long line of attorneys waiting to prosecute. Too bad liability attorneys only go after people and companies who manufacture tangible products and not people and companies that churn out financial products.
Your editorial brings up some good points about skepticism and attempting to model a system that cannot be tested outside the model. If you leave out one variable, the whole model is invalid. The problems with models used by financial institutions was that they neglected the role politics played. If, in fact, they had been able to use historical data rather than forced political agendas in assessing risk in mortgage obligations, their risk could have been moderated just as it was for decades. Additionally, one could argue that political agendas skew climate modeling as well.
Your recent editorial is right on target. Computer output depends on computer input. I remember two astrophysicists at the National AGU meeting in 1976 telling me that stellar explosions are always symmetric, completely mixing up material from various stellar layers. They also insisted material ejected from a supernova cannot form a planetary system orbiting a supernova remnant. When I inquired how they knew stars always explode symmetrically, they replied they had modeled stellar explosions on computers. I asked if they considered the spin of the star before it explodes. They replied, “No, our computer can’t handle spins, so we threw it out.”
The Hubble Space telescope eventually helped scientists discover that stars routinely explode axially, like SN1987A. The first planetary systems found outside our solar system had three Earth-like planets orbiting a pulsar, the core of a supernova. We now know that the astrophysicists themselves were riding on an ironrich object that orbits near the core of a supernova that exploded about 5 billion years ago and ejected all of the material that now orbits our Sun.
Oliver K. Manuel
The deadly grind
I found your Berke on Safety column regarding grinding wheels very informative (“The case of the exploding grinder,” Feb. 6). I wish it would have included a recommendation to always use a full face shield that covers both face and throat area when grinding.
My medical rescue squad unit was recently dispatched to an industrial site for a nonresponsive man. Upon arrival, we saw a man laying unconscious and not breathing. Unknown to us, he was well away from where he had been working and no one could give any information about what happened. We conducted emergency lifesaving, but all our efforts failed.
Later, we were able to put the accident together. It turns out a grinding wheel on a hand grinder failed, shattering during a grinding operation. A piece of the wheel struck the worker in the throat crushing his windpipe. In a heroic effort to save his own life, the workman tried to cut his own throat and open his airway but failed. This could have been prevented if the workman had worn a full-face mask.
It shouldn’t take a Sherlock Holmes to figure out what went wrong with the grinder. The worker never should have put a 7-in. wheel on it in the first place. The notice on the grinder says the largest wheel should be 5 in. in diameter. When the manufacturer believes it is important to put a tag or warning on a piece of equipment, you should take the time to read it.
I have bought and driven trucks for the last 30 years, so my observations aren’t scientific or comprehensive. However, I can say unequivocally that I have never seen the same truck with the same engine get better gas mileage with an automatic transmission compared to the manual transmission option. Of course I recognize technology could be closing this gap. And there are few trucks sold with manual transmissions anymore. So it does seem odd that automatic shifting, which theoretically could be optimized, still carries a mileage hit.
So why so few manual transmissions? Convenience and laziness. But in drivers’ defense, if you live in a large metro area, the convenience is well worth the small (1 to 3 mpg in my experience) decrease in gas mileage stemming from an automatic transmission.
Off the Web
The recent article, “Manufacturing Technique Produces Parts Lighter and Stronger Than Aluminum,” (Feb. 5) needs some points clarified.
1. What the article refers to as MAG for Magnesium Injection Molding is Thixomolding. And Thixomolding and Thixomolded are registered trademarks of Thixomat Inc., Ann Arbor, Mich. We guard these marks zealously to prevent their losing their trademark status. It appears Phillips is separating itself from the Thixomolding process by establishing its own identity and that certainly is a viable marketing strategy. However, we have a legitimate concern about confusion in the marketplace as to whether the basic MAG process is separate and distinct from Thixomolding.
2. The Thixomolding process was first patented by Thixomat in the very late 80s and has grown since that time to encompass 50 licensees worldwide using more that 350 Thixomolding machines in 12 countries on three continents. Phillips is one of those licensees and has been successful with our process, including winning numerous awards. We are proud of how they have marketed and mastered the process.
3. Thixomolding itself is defined as the injection molding of magnesium alloys using a process similar to plastic injection molding except you mold magnesium alloys. It should not be confused with MIM or PIM processes.
4. Our primary concern is protecting our Thixomolding and Thixomolded trademarks. We are most zealous in defending these marks and get particularly nervous when we seen them used in stories such as this one without being capitalized and without the notification that these marks are registered. As one learns from aspirin, elevator, and other terms now part of the generic world, if you do not protect your mark, you will lose it.
Herbert E. Pritzker
Marketing Director and General