I research public markets and write up what I find.

Most pieces start with a big market move and the explanation that forms around it. I check that explanation against the data underneath it and write up whether it holds. When the research needs software, I build it.

Read the latest post

Latest post

All writing

Seat-Based Software Grows With Hiring, and Hiring Has Stopped

The software slowdown is real, and so are the loan losses. The explanation most people have settled on, that AI is replacing the products, doesn't fit the data nearly as well as a simpler one: seat-based software tracks white-collar hiring, and hiring stopped.

26 OCT 20266 MIN READ

Projects

Software I’ve built for this research, with notes on how each one works.

See projects

How I work

  1. Group the day’s price moves by the driver behind them, usually a macro event or a news story.
  2. For each driver, state what the market is assuming.
  3. Test that assumption against second-order indicators to tell correct pricing apart from a misread of the driver’s details.
  4. Shortlist only the misreads.
  5. Check each shortlisted company for risks of its own, like earnings, debt or litigation, then set the entry.