Friday, April 07, 2006

My Renowned Statisticsologist Says . . .

Here's the analysis I just got back from My Renowned Statisticsologist:

(A) Well, first calculate the probably of one acquaintance having an accidental death in the family. This — in the long run — is simply the number of families that have accidental deaths in the average week divided by the number of families (this is in the world, in the US, in your state — whichever you think is most relevant).

Now, two calculate the chances of two in the same week, you need to define your terms a bit more explicitly. (1) Do you mean in one particular week (e.g., what's the probability of two such occurences in the week of March 3-9)? (2) Or do you mean at some point in your life two will happen in a week? (3) Or do you mean that given one happened, what's the probability of another happening within 7 days?

(3) is easiest because we can make a reasonable assumption about independence because we will clearly be talking about two separate events. There might be some correlation, but it's probably small. Then the answer would simply be whatever you got for (A).

(1) is next easiest. If we assume independence (a big assumption) then you just square (A). If we think there is some positive correlation, i.e., acquaintances' family members might often travel together in the same car, then this will be closer to just plain old (A). Or, if accidental death includes things like death by earthquake, then there's certainly some positive correlation. If you really mean, a safe dropping on someone's head or falling down the stairs, then these are probably close to independent and your answer is (A)^2

(2) This is the hardest. Since the chance is so small you can approximate it by 1 - ((1 - "1")^N) where "1" is the answer you get in the above paragraph and N is the # of weeks you live in a year.

Did you get all that, Brother/Sister? Well, just in case you didn't, here's the upshot: the odds are ASTRONOMICAL that my two Stenos' family members would suffer near-simultaneous mortal "accidents." Like a gajillion-and-a-half to one. So I'm going to proceed on a double-homicide theory.

And Brothers and Sisters, please do let this insider's access I've given you into my management of this crisis inform your understanding of how I intend to govern the planet, once I take over. A Benevolent and Wise World Leader does not act willy-nilly — he consults with trusted and knowledgeable advisors and makes rational policy conclusions that incorporate and reflect their analytical expertise.

That, people, is how you govern.

Whether or not I decide to formalize the position of Renowned Statisticsologist — er, Statistician — into my regime structure, you can bet I will be relying heavily on this aforequoted gentleman's number-crunching insights well into the future.

No comments: