Where code meets finance

There's a lot of overlap between writing good software and doing good financial work. Both reward the same instincts.

Models are just programs

A spreadsheet model and a script are the same idea wearing different clothes: inputs, logic, outputs. The difference is that code makes the logic explicit and testable. Once I started treating my models like software — version-controlled, documented, checked against edge cases — they got a lot more trustworthy.

Good assumptions beat clever math

The hardest part of a model is rarely the math. It's the assumptions. Naming them, writing them down, and stress-testing them is where the real work is — same as handling the edge cases that break a program.

Make the assumptions loud. The math will follow.

That's the thread I want to pull on across this site: taking the discipline of engineering and pointing it at problems with numbers attached.