About 100 years ago, Sir Francis Galton took an ox into a marketplace and asked lots of people to guess how many pounds the ox weighed. The individual guesses came from a diverse sampling of people, who weren't educated, and it was Galton's intention to demonstrate that the guesses of a few people would be close to the actual weight of the ox, while most of them would be way off the mark.

The guesses did range far and wide, but what he found, surprisingly, was that when he took the average of all the guesses, they were off by a single pound. And while it would be easy to bill that as a happy coincidence, Galton's experience was not an isolated phenomenon. It's come to be known as the "wisdom of crowds", and lots of examples of this are documented in James Surowiecki's book of the same name.

I don't know about you, but the premise of the wisdom of crowds concept originally made me gnash my teeth. From both intro Psych classes and bitter experience, I've learned that group dynamics don't tend to result in more intelligent conclusions being reached. As Tommy Lee Jones put it in Men in Black: "A person is smart; people are dumb."

But there's a critical distinction to point out here: Galton had each person guess how much the ox weighed individually. This is important. If Galton had gathered a large group of people together and said, "Now, all of you, discuss amongst yourselves how much this ox weighs until you all reach a consensus", well, that would have been an entirely different situation. Chances are good that, given the opportunity to argue and bicker with each other, that crowd of people would have come up with a really bad guess as to the ox's weight.

Instead, everyone made their own guess, based on their own personal judgment, without being swayed by anyone else. As long as each person is acting independently, and you aggregate their knowledge, then the wisdom of crowds works.

Shortly after reading the book, I got to wondering if this could be applied to weather. The old joke is that the weatherman is one of the few jobs where you can be wrong 90% of the time and maintain steady employment. And, of course, that we have lots of expensive doo-hickeys called satellites flying around the Earth, but we can't tell when it's going to rain as well as a herd of cows can.

What if you asked everyone for a guess about tomorrow's weather? Ask them for the high and low temperature, and a guess about conditions through the day. If you averaged all those guesses together, how accurate would it be?

I was interested in getting an answer to this question, so I clobbered together a website called Weather Collective (with some help from my girlfriend) that aims to figure out if the crowd can guess tomorrow's weather. For now, we've set it up with about two dozen cities in the U.S., just to see what happens. If you're interesting in participating, please feel free to register and start making guesses along with us.

Like most new websites, getting all the features fully in place is going to be an iterative process, so please forgive any bugs you encounter. All feedback is welcome. Thanks!