I had a friend forward me a copy of the Victoria Camera Club's summer newsletter, which had winning images from a year-end contest. The "BMX Racers" image on page 13 interested me because its subjects are blurred, giving a sense of motion. I've seen this before, but never really tried to do it myself.
My first attempt was our Christmas tree. My thought was that a photo of a "dancing" Christmas tree might be amusing, so I set up in the living room and did some experimenting. My basic method was to set the timer on the camera for 10 seconds, giving me enough time to jump behind the tree and start shaking as soon as the shutter opened. I think I tried just about everything. Living room curtains open, closed; front flash, rear flash, no flash; different amounts and speeds of tree shaking; unplugging the tree lights half way through... Fortunately no ornaments were broken in the process, but I never really ended up with an image that satisfied me. Thinking back, this may be because my exposures were too long (20 to 30 seconds) which made the tree look just plain blurry. If I had gone with a shorter exposure (say, 2 seconds), I could have caught a single movement of the tree, rather than just a bunch of random convulsions. Looking back at the BMX racers photo, it is clear that each rider is moving on a fairly regular path.
My second attempt was of my girlfriend riding her rollers in the kitchen. Although slightly better than the Christmas Tree photo, I think it still suffers a bit from too long of an exposure (which in this case, was 1/2 second). Although this time I was constrained by the amount of light indoors. A few more criticisms are that I would like for there to be more static content in the photo, something fixed to contrast with the motion (other than the damn stove). I would also change the lighting a bit. We have a grid of 6 overhead lights in the kitchen, and I think slightly less uniformity to the lighting would give stronger shadows, adding depth and texture.
P.S. This post breaks me out of my 1 photo deficit.