I’ve seen this video, by NPR, in a couple of places recently:

Kottke was one of them, not sure of the other. It’s about the fact that under all sorts of different tests, people have been shown consistently to walk, swim or drive in spiral patterns if they are blindfolded and ordered to move straight ahead. Now… clearly some scientists are genuinely puzzled by this, so it seems unlikely that there is an obvious solution. But I can’t help thinking: when you blindfold someone and order them to walk in a straight line, surely the very likeliest outcome is that they go in a spiral? To go in a spiral, they just have to almost, but not quite, walk in a straight line, and do that consistently. And the likelihood of walking in a manner that is almost, but not quite, straight, is vastly more probable than that of walking perfectly straight? There are millions of ways that you can be slightly off, and only one way that you can be perfect, hence it’s almost certain that you will spiral.

At least, that’s what I reckon. I wonder if the curvature of these spirals is consistent? That would be interesting. Anyway.