Assembling the Machine

Extend your Foundations page: the books and their goals; each book's target allocation and benchmark; the position cap; the rebalance trigger and its materiality floor; the storm paragraph. If a rule isn't written, it doesn't exist — it's a mood scheduled for negotiation. One page still suffices. It always did.

The machine runs on inputs: new money lands by rule — this share to the core, this to the near-goal book, this trickle to experiments — automated where possible, decided nowhere in the moment. You built this in Foundations as dollar-cost averaging; now it's the fuel line of a larger engine. Money that arrives pre-routed never meets your moods.

On schedule: check drift against bands, rebalance what's material — sells first, then buys — and grade each book against its benchmark, risk-adjusted, drawdowns included. Then stop. That's the entire recurring workload, an hour a quarter at most. Everything else the market screams at you between maintenance windows is, by construction, not your job.

Once a year, the allocator convenes: did life change the goals? Did any book earn a bigger mandate or prove it deserves a smaller one? Does the page need amending — in calm, in writing, effective next quarter? Rules may change; they may never be BROKEN mid-storm. That distinction is the machine's warranty card.

Notice what you've assembled: policy targets, hard caps, scheduled rebalancing, materiality floors, honest benchmarks, drawdown rules — it's the same skeleton inside pension funds, endowments, and our own engine, where every one of these rules runs as code. The masses are sold picks because picks are exciting. You now own the boring architecture that actually decides outcomes. Run it.