Tuesday, October 26, 2010

Lies, Damn Lies, and This

What are the chances that any two people in a group of thirty would share a birthday? Given the 365-day years we have here on planet earth, the chances must be pretty low, right? Wrong!

Using careful calculations we encounter a counter-intuitive conclusion: in a group of 30 people, the chances that two people share a birthday is above 20-, 50- and even 60-percent. On face value, not many of us would probably think the chances that high. This shows there is actually nothing remarkable or special or spooky about two people sharing a birthday, considering that cold calculation indicates the likelihood being more than a coin-toss.
This sounds like a good way to win a free drink at a bar, as are so many sleights of hand and mind that, in the end, result from the way our perceptions go wrong. Statistics can be fun -- often they are not, frankly, but they are very useful:
Empowering ourselves with numbers might seem strange, until we recall how statistics can destroy the pretentions of charlatans or miraculous happenings. Indifferent in itself, statistics displays information anyone is welcome to assess.
Reality is too vast and interesting for us at times, so we have tools.


[First paragraph corrected thanks to "Anonymous"]

3 comments:

Paul Sunstone said...

It seems humans have some instincts for number that are largely accurate and sound, and yet some other instincts for number that are largely inaccurate and misleading. That is, we are a bit mixed when it comes to what our instincts for "math" tell us is likely or correct in a given situation. Makes me want to go back and blame all my low grades in high school math on evolution.

Dale said...

Paul, you would be right to do so! In general, our minds seem a poor fit for the kinds of math we've nevertheless managed to eke out.

Paul Sunstone said...

That reminds me: I was reading a while back that chimps outdo us in some cognitive functions. I can't recall exactly what those functions were, but I seem to remember one having to do with statistical reasoning. If so, that's quite amazing.