Decisions
What choice would move the problem if you made it now?
Good decisions turn uncertainty into chosen action. They start with a real problem, sharpen through good questions, and improve through a decision journal.
Decision quality shows up in reality, not in how confident the decision felt.
The Spine
- Problem solving - frame the problem.
- Algorithms - choose method.
- Problem domains - route context.
- Decision journal - record proof.
The Protocol
Every non-trivial decision runs seven phases:
- Classify the door — reversible means act fast and learn cheap; one-way means verify slowly; compound means the decision shapes future decisions, so design it to compound.
- Verify state — list what you know from evidence and what you don't. No claim without proof; absence of evidence is not evidence.
- Name at least three options — binary choices hide better choices, and "do nothing" always counts as one.
- Chain the reasoning — per option: the evidence for it, the hidden assumption behind it, and the inversion that would make it wrong.
- Pick with honest conviction — high only with verified evidence. If you cannot name what would flip the call, you are not ready to make it.
- Write the receipt — record the pick, options, inversion, prediction, and kill signal in the journal so future-you can audit the process, not just the result.
- Schedule the review — a date and the evidence you will check. A prediction you cannot track was never a real prediction.
If you cannot name the choice, you are still questioning. If you cannot name the proof, you are guessing.
Worked instance: the country analysis prompt runs this protocol at country-platform scale — evidence gates, sensitivity, counter-case, and a decision brief that feeds the journal.
Failure modes: fake decisions change nothing, vague decisions name nothing, unmeasured decisions prove nothing, and fear decisions avoid regret instead of serving the value setpoint.
Context
- Navigation gives decisions their value, belief, and control frame.
- Meetings are where group choices become clear or expensive.
- Tech stack decisions apply the loop to technology choices.
Zoom Out
Decision quality compounds when each choice leaves a better record for the next one. The goal is not certainty. The goal is a better loop.
Questions
What decision are you avoiding by gathering more information?
- If decisions are judged on process, how would you rate your current process?
- Which decisions run on values, and which run on fear?