'This never happens,' chirped my wife's doctor as he declared our second set of twins.

'If it's going to happen, it will happen to your family,' noted my child's doctor after emergency surgery to repair a routine tonsillectomy.

'I mean this with all due respect,' continued the family dentist, 'but your family is what we in medicine consider a 'statistical outlier.'  Somebody has to make up the tail of the bell curve and you're it.'

Three data points make a trend I suppose. Hearing these cheeky remarks were funny the first time or two.  Now, we budget for the tail of the curve.  The bell curve may apply to the majority, yet, by definition, there must be an outlier; two or three standard deviations from 'normal.'

The business world will often round up or down in getting the last few percentage points because the effort for precision exceeds the value found in complete accuracy.  But that doesn't work with people.  Or at least it shouldn't. People aren't 'rounding errors.' The total must add up to 100%, so when the doctor tells us that the cause of the emergency surgery only happens in 1% of cases, discussions of the bell curve shift toward questions of operating room cleanliness and recovery times.  I don't care about statistics -- I care about my child. 

And the doctor did too -- when I woke him up at 3:45 a.m. The doctor knows the statistics because he knows the context.  It's his job to understand the outliers, why they happen and how to avoid them.  He also knows my child because he knows our context, having pulled the tonsils of all of our children.  It is in this context that adjusts his approach to the data.

Statistics are helpful for context, yet there's a human in each percentage point.  My life insurance provider disagrees with me, but they don't count [some puns are too easy].

If we are indeed in the 1% of outlying cases, then please excuse me, I must buy a lottery ticket.