top of page
  • Writer's pictureLeah Zani

The Guessing Game

Reviews of Bomb Children are coming in, and this review by Erin Lin stood out to me as exceptionally well-written and informative. Lin is too humble to mention that she is also an expert in the social study of military waste, so I want to highlight her work here:

Lin is an Assistant Professor of Political Science at Ohio State University, where she specializes in the study of postconflict reconstruction and legacies of war in Southeast Asia. She’s currently working on her first book, How War Changes Land, and Chapter One won the best paper award from the Midwest Political Science Association. The book is an ethnographic and statistical study of the covert US bombing of Cambodia, a sister conflict to the war I studied in Laos. I look forward to reading her book in its entirety!

Lin has been conducting research in Cambodia for twelve years in partnership with the World Bank, the Center for Khmer Studies, and MAG. Among other things, she uses machine learning and statistical analysis to better gather evidence of military waste, such as bomb craters and chemical leakages. A recent press release for her machine learning project announces that this new method increases accurate detection by more than 160 percent over standard methods.

The new models generated by Lin’s machine learning methods, in combination with satellite imagery and declassified US military records, indicate that 44-50% of the bombs in her study area remained unexploded. This accords with my own suspicions about failure rates for air warfare: The Lao government tends to tout a 30% failure rate, but I have been unable to track down a verified source for that statistic. Failure rates are so specific to terrain, weather, type of ordnance, etc. that I doubt it is possible to create a standard rate for a nine-year conflict. In some of the villages I studied forty years after the war ended, people were still living with ten or more live bombs per square meter—to me, this implies a failure rate much higher than 30%.

Like Lin’s machine, I had to learn to identify craters in the field (like the ones in the images above). I created a guessing game with one of my clearance interlocutors: One of us would point at a hole in the ground and the other had to guess if it was a crater, and back up their guess with evidence. I learned to identify the small lip of debris that surrounds true craters, persisting even decades after an explosion; debris from shrapnel and broken bomb cases; the relative width and depth of craters for different kinds of ordnance; clusters of nearby holes that might indicate a strike pattern; and evidence of surrounding warfare, like bullets in a field.

You can learn more about Lin’s work at her website.

32 views0 comments
bottom of page