I noticed that my little machine is more accurate than the predictions of groups of economists, (for example, last week my model predicted that new unemployment claims would jump by 2500, whereas a the economists thought that they would decline by 1,500. They were off by 7,000; I was off by 1,500. They moved in the wrong direction; I moved in the right direction. Not to brag.).
First the numbers:
Jobless claims rose by 4,000 for a second week to reach 372,000 in the period ended Aug. 18, Labor Department figures showed today in Washington. The median forecast of 41 economists surveyed by Bloomberg called for 365,000. The four-week moving average, a less volatile measure, increased to 368,000.
Here’s my secret. I have made a machine to predict the initial unemployment claims. Since I’m not especially handy with tools, I’ve constructed it with parts I found lying around the garage. First I took my old set of roller skates, from when I was a little girl, the kind that fits over your shoes and adjusts with a key. You know the kind? They are the new base of a special wheel for my pet hamsters, Barney and Frank. Imagine a hamster wheel on a simple wooden platform, which is affixed to the tops of the roller skates. The wheel can spin easily, just like a ferris wheel.
Once a week, on Thursday mornings, I take Barney and Frank out of their stationary cage and put them onto Aggie’s Complicated Mathematical Model™., which is on the floor of the garage. Then I gently push the platform, just enough to get the roller skates moving, which usually causes Barney and Frank to run on the wheel. The direction of the spin determines whether the unemployment numbers go up or down. The intensity of the spin determines by how much. Sometimes they kind of freak out, like the Keystone Kops, and bump into each other. That means little or no change.
And sometimes, no matter how much I spin the roller skates, I can’t wake them up. They are fat and lazy and prefer to be left alone. On those occasions I take out my other secret ingredient: Cynicism. I make the assumption that things are getting worse, or at best bumping along the bottom.
Now, I realize that not everyone in Bloodthirstan has hamsters. But you can do something similar with goldfish. Tap on the tank, or make faces or something, and create predictions on which way they flee – back, forth, up, down, like particles in physics. For a new puppy, count the spins as he chases his tail. Be creative!
I am hoping that we can level out the bumps in my model by having a lot of participants each week. We can average our numbers, just like the pros.