In 1945, the US Secretary of War's office "estimated that conquering Japan would cost 1.7 to 4 million American casualties, including 400,000 to 800,000 fatalities, and five to ten million Japanese fatalities. The key assumption was large-scale participation by civilians in the defense of Japan."

"Officially ... There were 96 'industrial' fatalities. during the construction of Hoover Dam. This figure does not include deaths by other means including heat, heart problems and curiously, pneumonia. Another estimate is 112 and that supposedly includes a longer time frame than just the construction period. Oddly the larger estimate also holds some irony. It begins the count on December 20, 1922 with J.G. Tierney a Bureau of Reclamation employee who was part of a geological survey and drowned when he fell from a barge. Exactly 13 years later, in 1935, his son Patrick W. Tierney, fell to his death from an intake tower."

We all take risks every minute of our lives. We can reduce risk -- not keeping a loaded firearm in a house with children, for instance -- or increase it.

When does risk become unacceptable? The two quotations above are examples of large-scale projects that had estimated deaths and injuries. Actuaries and "risk managers" deal with it all the time as numbers; the people who are injured or die have more personal stakes.

What actually is an acceptable level of risk and how do you determine it for yourself?