Stats Made Easy

Practical Tools for Effective Experimentation

Sunday, August 05, 2007

The challenge of dealing with statistical anomalies such as bridge collapses

Thursday morning seemed surreal on my dawn commute to the shaken city of Minneapolis the morning after their downtown’s entryway collapsed catastrophically. Drivers darted this this way and that, seeking alternate routes. As I walked in from the parking lot, a helicopter hovered directly overhead our building taking pictures of the I35W bridge remains. Cartoonist Marshall Ramsey captured all of our anxieties nowadays by picturing a span held up by question marks. I just hold my breath and speed up until crossing over and then breathe a sigh of relief!

The really important question is the cause of this precipitous breakdown in this vital structure. CNN’s widely seen video seemingly pinned the south end as the point of failure, but the northern end of I-35W bridge is now focus of probe. I heard one expert on the engineering of bridges say we should not be overly concerned about another bridge failure anytime soon because this one is an “anomaly”!

When the consequences are trivial, it is easy to dismiss events of such extremely low probability. However, in an extremely high-impact case like this, we want to know the risks of it happening again – assuming no intervention will be accomplished any time soon. This weekend I saw the August issue of The American Statistician magazine, which features reviews of a book titled The Black Swan: Impact of the Highly Improbable by Nassim Nicholas Taleb. According to Taleb's home page named after his first book Fooled by Randomness /, Taleb’s goal in life is “…teasing people who take themselves & the quality of their knowledge too seriously & those who don’t have the guts to sometimes say: 'I don’t know'...."

Taleb borrowed the black swan concept from philosopher Karl Popper. Before these birds were found prevalent in Australia (beautifully detailed by Black Swan wines ), they were thought to be extremely rare by the Western world. Taleb says that it may be emotionally satisfying, but we accomplish little by trying to explain catastrophes such as the 9/11 attack, and fool ourselves by quantifying their risk of re-occurrence. (I wonder if this serves as an example -- the five threat levels predicted by the US department of Homeland Security, Low = Green; Guarded = Blue; Elevated = Yellow (current level); High = Orange; Severe = Red.)

Taleb obviously gets a kick out of tweaking the noses of pontificators purporting predictive powers on the next improbable high-impact event. However, I am studying with great interest all ongoing reports from the investigation of the I35 bridge collapse just down the road from my Stat-Ease office, not only for it being so proximate, but also for my peace of mind in having to constantly cross other spans over our region’s rivers and highways. I look forward to the day when our bridge gets rebuilt and perhaps we can again enjoy swans of any sort swimming along the banks of the Mississippi in downtown Minneapolis.


Post a Comment

<< Home