sporter

I'd love to see your take on the risk analysis numbers. I'm not even a math/econ/statistics geek and the flaw in the logic is obvious.

If you spend $1 billion on safety (generically) and one person dies then you didn't save 1 life for $1 billion. The number of lives saved is difference between the number that would have been lost had you not spent the $1 billion and number that actually did (or will) die. So if 101 people would have died, you saved 100 lives for $1 billion, or $10 million per life saved.

Blue Sun

The article in question pointed out that over the past 100 years, terrorism has been a rather minor problem overall. Anybody in the year 2000 trying to predict deaths from terrorism in 2001, not to mention the cost to our nation, would have been way off the mark. In an hour and a half, 19 men armed with boxcutters killed 3,000 people and set off an economic crisis that cost, by conservative estimate, a half a trillion dollars lost to the economy. If we factor in the wars in Afghanistan and Iraq that have followed the attack, that cost is over 7,000 lives and upwards of 2 trillion dollars, and the cost, both in lives and treasure, continues to rise today.

After all, it would have been perfectly logical for risk analysts in 1944 to predict, based on a 100-year history, that the mean and median numbers of people who would die in a nuclear holocaust in 1945 would be 0.

The world is changing ever more quickly, and the risk analyst's favorite tool of analyzing historic trends and extrapolating them into the future can not be relied upon any more.

Read more...

DC

Oh man, read the comments on that first Delta post. I was dying.

http://blog.delta.com

John K. Lunde

To comment number 2, you're talking about a phenomenon known to many people today as the "Black Swan" thanks to Nassim Taleb. You should check out his books if you're interested in such things, as they're a pretty fascinating thought process on forecasting.

To sum things up, it's what you don't know you don't know that will kill you in trying to forecast the future. Or, to use another popular phrase from investing, past performance is not necessarily indicative of future results.

sporter

I'd love to see your take on the risk analysis numbers. I'm not even a math/econ/statistics geek and the flaw in the logic is obvious.

If you spend $1 billion on safety (generically) and one person dies then you didn't save 1 life for $1 billion. The number of lives saved is difference between the number that would have been lost had you not spent the $1 billion and number that actually did (or will) die. So if 101 people would have died, you saved 100 lives for $1 billion, or $10 million per life saved.

Blue Sun

The article in question pointed out that over the past 100 years, terrorism has been a rather minor problem overall. Anybody in the year 2000 trying to predict deaths from terrorism in 2001, not to mention the cost to our nation, would have been way off the mark. In an hour and a half, 19 men armed with boxcutters killed 3,000 people and set off an economic crisis that cost, by conservative estimate, a half a trillion dollars lost to the economy. If we factor in the wars in Afghanistan and Iraq that have followed the attack, that cost is over 7,000 lives and upwards of 2 trillion dollars, and the cost, both in lives and treasure, continues to rise today.

After all, it would have been perfectly logical for risk analysts in 1944 to predict, based on a 100-year history, that the mean and median numbers of people who would die in a nuclear holocaust in 1945 would be 0.

The world is changing ever more quickly, and the risk analyst's favorite tool of analyzing historic trends and extrapolating them into the future can not be relied upon any more.

Read more...

DC

Oh man, read the comments on that first Delta post. I was dying.

http://blog.delta.com

John K. Lunde

To comment number 2, you're talking about a phenomenon known to many people today as the "Black Swan" thanks to Nassim Taleb. You should check out his books if you're interested in such things, as they're a pretty fascinating thought process on forecasting.

To sum things up, it's what you don't know you don't know that will kill you in trying to forecast the future. Or, to use another popular phrase from investing, past performance is not necessarily indicative of future results.