One thing I’ve got to become better at: documenting what I do while I’m doing it (why did you make that formula that way?) so I can piece together the clues several months later when I’m actually writing the report. I’m in the middle of writing something up for work I did four months ago and my current writeup sounds horribly bloodless and un-colorful when I remember sweating out the work and fighting over what I thought was a clever conceit but which ended up being just extra work, misdirected. The final spreadsheet was clever and had smart tweaks to put in random disturbances; it was also immense, but that’s to be expected when you’re putting together data for every hour over five years for twelve-plus units.
I like the idea of forecasting given some simple constraints; I have colleagues who specialize in that sort of thing and it makes me happy to think of what you’re physically doing: short lever, long arm, and no one to say you’re doing it wrong unless of course the numbers are completely unreasonable. You know what they say about past performance being no predictor of the future? We make hypotheses and the educated guess is as good as you’re going to get, assuming you’re not hung up on trying to get it perfect. Throw in enough variables and you could start to predict almost anything, it feels like.
Life is unpredictable, and frequently messy besides. There’s no hard data on projecting the cost of diapers, though I’m sure diapers futures exist as a market. There’s other variables: when is your kid getting toilet trained, will they get sick, what about using strange bathrooms, how will that work. You can make educated guesses about all of them but the regulations in that market are a lot less stringent than for public utilities. You don’t need to have perfect predictions for everything (although lottery numbers would be a good start) as long as you understand the limitations of the art. We won’t — can’t — be perfect about everything, and that kind of control is hard to release at times.