I have noticed that I understand something roughly twice as well after I have tried to write it down. The spreadsheet gets me to an answer; the paragraph is what tells me whether the answer holds up.
So this is the place for that.
What goes here
Three kinds of things, mostly:
- Company and market notes. What I am reading, what I think it means, and — when I get it wrong — what I missed. I would rather have a public record of a bad call than a private memory of a good one.
- Things I learn building StrideSignal. Cleaning 38,712 rows of para track results taught me more about data than any class has. Most of that lesson was about how messy real data is before it is useful.
- Methodology. How a number was actually built. Every analysis has a dozen small choices buried in it, and those choices usually matter more than the conclusion.
The standard I am holding myself to
Show the work. If I claim a company is mispriced, the assumptions should be visible and you should be able to disagree with them specifically. If I adjust for inflation, I should say which index and why.
That is the whole idea. Numbers that can be checked, and a record of what I thought at the time.