Wednesday, August 18, 2004

Randomness?

A few years ago, I heard a story about a university experiment of pattern analysis on random numbers, which I remembered something like this:

Some university has two programs running simultaneously on a mainframe. One perpetually generates a stream of random numbers – mathematically random, not like a low-level RND pseudorandom function found in many programming languages. The other one perpetually analyzes the stream for repetitive patterns. Short repetitions are relatively common, but every so often, a repetition occurs that is statistically extremely improbable. Analysis shows that an even more improbable percentage of these occurences happened within an hour of some major world event – earthquake, war or terrorist attack, death of a world leader, etc.

The story has many of the hallmarks of an urban legend, especially the whole "powerful knowledge They don't want you to know" feel of it. I had thought about it a couple of times, but had mostly forgotten about it... until yesterday, when I read this post.

I commented on it, and didn't get any direct answers, but there were a couple of (confrontational sounding) comments from one reader trying to get me to defend the claims of the story. I explained that I was not defending it, I was just wondering if anyone had heard of it and whether or not it was a hoax.

As I was trying to properly word my response, I thought it would be a good idea to find a reference to explain the meaning of "mathematical randomness" as I understood it. When I googled some search terms, the FIRST link listed both explains the concept, and has a link to the EXACT EXPERIMENT I was asking about in the first place - noosphere.princeton.edu.

How random is that?!?

The experiment is not quite what I remember. The premise of the experimenters sounds like new-age-global-consciousness nonsense, but the experiment itself is interesting (just click on the data access link, and there are several options available for viewing the data itself).

No comments: