Twenty-nine years ago, the Department of the Air Force paid the Research Triangle Institute to build a better way to estimate how often space boosters blow up. The result, a document now released by the Department of War under the PURSUE archive, landed on September 10, 1996. Its title is dense: DOW-UAP-D48. Its subject is narrow: modeling unlikely space-booster failures in risk calculations. But the report’s existence tells a broader story about how the military learns from its own catastrophes.
The document was prepared for the 45th Space Wing and 30th Space Wing Safety Offices. Those wings manage Cape Canaveral and Vandenberg Air Force Base, respectively. Between them, they handle the bulk of U.S. military and intelligence satellite launches. When a Titan IV or a Delta II fails on the pad or breaks apart over the Atlantic, those safety offices are the ones who figure out what went wrong and how to stop it from happening again. The Research Triangle Institute report was their attempt to systematize that process.
James A. Ward, Jr. and Robert M. Montgomery, both of the institute’s Center for Aerospace Technology Launch Systems Safety Department, wrote the report. They catalogued historical launch failure modes. They proposed novel modeling techniques to improve risk calculations. The goal was straightforward: identify the risks that everybody already knew about, then model the ones that were so improbable nobody had bothered to quantify them before. Unlikely failures kill people and destroy hardware precisely because they are unexpected. Ward and Montgomery aimed to make the unexpected predictable.
The timing matters. 1996 sat in the middle of a decade that saw several high-profile launch failures. A Titan IV exploded over the Pacific in 1993. A Delta III failed in 1998. The Air Force was spending billions on the Evolved Expendable Launch Vehicle program, which would eventually produce the Delta IV and Atlas V. Getting risk calculations wrong meant losing payloads that cost more than the rockets themselves. Getting them wrong on a crewed vehicle meant losing astronauts. The report was part of a broader push to bring actuarial rigor to an industry that had historically relied on engineering judgment and crossed fingers.
The Department of War’s release of this document under the PURSUE archive offers a window into that effort. The report itself runs 21.5 megabytes, dense with tables and equations. Its summary is spare, revealing little beyond the title and purpose. But the fact that the Air Force commissioned this work at all says something about how the military views space. It is not a frontier to be explored. It is an operating environment to be managed. Failure is not an adventure. It is a data point.
Where this leads is predictable. The modeling techniques Ward and Montgomery developed in 1996 have likely been refined, expanded, and embedded into the safety processes that govern every military launch today. The 45th and 30th Space Wings still exist. They still launch rockets. They still use risk calculations to decide whether to press the button or scrub the mission. The report is a fossil of that system’s evolution. It shows the moment when the Air Force decided that gut instinct was not enough, that the improbable needed to be modeled, that safety required math.
The document is dry. It is technical. It contains no drama. But it is also a record of a bureaucracy trying to get smarter about danger. That is not a story about heroism. It is a story about process. And process, done right, is what keeps rockets from falling on people’s houses.
