If your income comfortably outruns your spending, a budget tells you what you already know. If it never will, no amount of categorizing transactions changes that fast enough to matter. Either way, the hour is better spent elsewhere.

I should admit I'm not entirely rational about this part. I've done the books for 5 companies, across France, China, and Hong Kong, and something about a ledger that balances to the cent still satisfies me in a way I can't fully justify. Most people find accounting boring. I find it the closest thing to proof I get in a day.

Placeholder — a ledger, or something from one of the 5 companies' books.

I keep one because I don't have steady income, and I'm carrying costs that currently belong to someone else and won't stay that way. That's a narrow, specific kind of uncertainty, and it's the only kind that makes this worth the attention. Most years I didn't need any of this. The year I did, I needed it precisely.

The honest version of that need never looked like a dashboard. It looked like a spreadsheet that kept getting asked to do one more thing, for years, until it was quietly running a 10-year survival estimate through formulas nobody should trust that much. So I built something that could actually hold the weight: import the real numbers instead of retyping them, keep categories straight across years in a way a spreadsheet can't, and forecast without pretending.

Placeholder — the spreadsheet, before any of this existed as software.

The forecasting part is the one I'm proudest of getting wrong first. I spent real hours trying to get a model to predict my future spending. It couldn't, not honestly. What works instead is almost embarrassingly plain: I say what I expect to spend, it tells me how wrong I've historically been about that, and shows me both. Best case, worst case, the date each one goes negative. No model pretending to know more than the data does.

Placeholder — the forecast itself: today's balance, the projected minimum, both cases.

It's since had to hold someone else's numbers too, not just mine. That's the only test that matters: does it still work when it isn't your own money anymore.

So no, this isn't for everyone. It's for the narrow band of people whose income doesn't let them stop paying attention, who'd rather see an honest range than a confident guess.