Before the Challenger disaster, engineers estimated that the space shuttle had a 1% probability of failure. Such a figure is seemingly low--few of us carry umbrellas with a 1% probability of rain--yet as America prepares to return to space, we must realize that Discovery's odds are not dramatically different than Challenger's.
In an incredibly complex system like the space shuttle, managed by a necessarily huge bureaucracy, parts can always fail. Some apparently obscure facts will not be passed all the way up the chain of command. It has been reported that there are 1,232 parts of the Discovery that, if even one breaks, can cause mission failure. If the probability of each of those different failures is as low as 1 in 100,000, there is still a 1.22% chance that something will go dreadfully wrong.
When, after only four launches, the National Aeronautics and Space Administration declared the space shuttle to be "fully operational," it encouraged our natural tendency to discount the hazards. Launches became too routine to be news. By the end of 1985, most television appearances of the orbiter happened in commercials. These ads, like Ford's for its Aerostar van, had various products comparing favorably to the engineering achievement of America's proven "space delivery truck." The small probability of failure had been fudged to zero.
The myth of the shuttle's perfection exploded with the Challenger. But public interpretation of the Rogers Commission report produced another familiar fantasy. The O-ring failure has been presented as inevitable. Half the respondents in an Associated Press poll this past July said their confidence in NASA was shaken by the shuttle explosion, and 30% of all respondents said they completely lacked confidence in the space agency.
Our past practice of crediting the successful flights to NASA's invincibility was an unrealistic denial of the risks. Our hindsight declaration of predestined failure is equally unrealistic. In this view, the Challenger had "no chance" while previous flights were "sure things" at the time. A closer examination of any of these missions shows that there were plenty of things that nearly went wrong with the successful ones, like O-ring failures on flight 51-C. And there were also problems that were anticipated but did not develop, like ice being drawn into the main engines during the Challenger's lift-off.
Test pilots understand their work too well to pretend that the chance of failure is zero and seldom have the luxury of looking back on disaster as fate. Tom Wolfe brilliantly captured their way of dealing with risk in "The Right Stuff." If a pilot has the "stuff," he is above and immune from the laws of probability.
These myths may be necessary coping mechanisms for the uncertainties in our lives, but they are no basis for a space program. Albert Einstein may have been right about God not playing dice with the universe, but we humans are forced to gamble as we explore. We hope that our scientists and engineers load the dice heavily in our favor. But we should not be fooled by success or quit after failure. Our astronauts should be admired for their bravery, not for their right stuff.
At all cost, we should not rewrite history after the Discovery: If it succeeds, as it most likely will, it will not turn the Challenger disaster into an aberration; if it fails, despite our best efforts, it will not prove that the first 24 missions were beginner's luck. Regardless of the outcome, we must accept the results, examine our loaded dice carefully, and then roll them again.