decision-making · diff
git:20260705.acb5571 to git:20260705.a28c4ce
14 added, 14 removed. Audit A to A.
---
name: Decision Making
- description: Runs a structured decision process for high-stakes choices — reversibility classification, weighted decision matrices, expected value with explicit probabilities, pre-mortems, second-order thinking, and 10/10/10 — and produces a documented recommendation plus a decision-journal entry. Use when someone asks "should I take this job or stay", "how do I decide between these two paths", "I keep going back and forth on a big decision", or "help me think through whether to relocate". Do NOT use for evaluating a specific large purchase like a house, car, or equipment — use big-purchase-decision instead.
+ description: Runs a structured decision process for high-stakes choices - reversibility classification, weighted decision matrices, expected value with explicit probabilities, pre-mortems, second-order thinking, and 10/10/10 - and produces a documented recommendation plus a decision-journal entry. Use when someone asks "should I take this job or stay", "how do I decide between these two paths", "I keep going back and forth on a big decision", or "help me think through whether to relocate". Do NOT use for evaluating a specific large purchase like a house, car, or equipment - use big-purchase-decision instead.
---
# Decision Making
- Most bad decisions are not caused by bad reasoning — they are caused by deciding the wrong question, rationalizing a pre-chosen favorite, or spending a week of analysis on a choice that could be undone in an afternoon. This process matches the depth of analysis to the stakes, forces criteria to be written before options are scored, and leaves a journal entry so judgment can be calibrated later.
+ Most bad decisions are not caused by bad reasoning - they are caused by deciding the wrong question, rationalizing a pre-chosen favorite, or spending a week of analysis on a choice that could be undone in an afternoon. This process matches the depth of analysis to the stakes, forces criteria to be written before options are scored, and leaves a journal entry so judgment can be calibrated later.
## Operating procedure
- Order matters: classification comes first because it decides whether the rest of the process runs at all, and criteria must be locked before any option is scored — scoring first invites rationalizing a favorite.
+ Order matters: classification comes first because it decides whether the rest of the process runs at all, and criteria must be locked before any option is scored - scoring first invites rationalizing a favorite.
### Step 1: Gather inputs
Collect before analyzing. Label any estimate the user cannot source as a guess.
- - The decision, stated in one sentence with a deadline. Default deadline if none exists: set one now — endless analysis is itself a costly choice.
+ - The decision, stated in one sentence with a deadline. Default deadline if none exists: set one now - endless analysis is itself a costly choice.
- The real options on the table. Push for at least three; a binary framing usually hides a third option (negotiate, delay, pilot, partial commitment).
- What happens if the user does nothing (the status quo is always an option and often the silently chosen one).
- Constraints that are truly fixed vs. merely assumed.
- - The user's emotional state (tired, rushed, angry, infatuated with one option) — record it; it belongs in the journal entry.
+ - The user's emotional state (tired, rushed, angry, infatuated with one option) - record it; it belongs in the journal entry.
### Step 2: Classify reversibility
- - Two-way door (reversible, low stakes): decide fast and move on. Rule of thumb: if undoing the choice costs less than a day of effort, spend minutes, not days. Decide with roughly 70% of the information you wish you had — waiting for 90% means you were slow.
- - One-way door (irreversible or high stakes — quitting a job, relocating, marriage, shutting down a product): run Steps 3-8 in full.
+ - Two-way door (reversible, low stakes): decide fast and move on. Rule of thumb: if undoing the choice costs less than a day of effort, spend minutes, not days. Decide with roughly 70% of the information you wish you had - waiting for 90% means you were slow.
+ - One-way door (irreversible or high stakes - quitting a job, relocating, marriage, shutting down a product): run Steps 3-8 in full.
Say which classification applies and why before proceeding.
### Step 3: Frame the real question
- State precisely what is being decided, what problem it solves, and whether that is the right problem. Surface hidden assumptions ("this assumes the promotion won't come through — is that confirmed?").
+ State precisely what is being decided, what problem it solves, and whether that is the right problem. Surface hidden assumptions ("this assumes the promotion won't come through - is that confirmed?").
### Step 4: Define weighted criteria, then score
List what matters and assign weights summing to 100% before evaluating any option. Separate must-haves (pass/fail gates) from weighted nice-to-haves. Then score each option 1-5 per criterion and compute weighted totals.
- Worked example — stay at current employer vs. join a Series B startup:
+ Worked example - stay at current employer vs. join a Series B startup:
```
Criterion Weight Stay Startup
Career growth 30% 2 5
Compensation 25% 4 3
Role/team fit 20% 3 4
Stability 15% 5 2
Location/lifestyle 10% 4 4
Weighted total 3.35 3.75
```
- Stay: (0.30×2)+(0.25×4)+(0.20×3)+(0.15×5)+(0.10×4) = 3.35. Startup: 3.75. A gap under ~0.3 is a tie — the matrix says the decision hinges on the weights, so stress-test them. Here the 0.40 gap favors the startup, but only because growth was weighted 30%; if the user quietly believes stability matters most, the weights were dishonest and must be redone.
+ Stay: (0.30×2)+(0.25×4)+(0.20×3)+(0.15×5)+(0.10×4) = 3.35. Startup: 3.75. A gap under ~0.3 is a tie - the matrix says the decision hinges on the weights, so stress-test them. Here the 0.40 gap favors the startup, but only because growth was weighted 30%; if the user quietly believes stability matters most, the weights were dishonest and must be redone.
### Step 5: Put explicit probabilities on uncertain payoffs
- Where an option's value depends on uncertain outcomes, compute expected value with stated probabilities instead of vibes. Example for the startup's equity over 4 years: 60% chance worth $0, 30% chance worth $40,000, 10% chance worth $400,000. EV = 0.6×0 + 0.3×40,000 + 0.1×400,000 = $52,000, about $13,000/year — compare that against the salary gap, not the fantasy outcome. Writing the probabilities down also exposes overconfidence: if the user cannot defend the 10%, lower it.
+ Where an option's value depends on uncertain outcomes, compute expected value with stated probabilities instead of vibes. Example for the startup's equity over 4 years: 60% chance worth $0, 30% chance worth $40,000, 10% chance worth $400,000. EV = 0.6×0 + 0.3×40,000 + 0.1×400,000 = $52,000, about $13,000/year - compare that against the salary gap, not the fantasy outcome. Writing the probabilities down also exposes overconfidence: if the user cannot defend the 10%, lower it.
### Step 6: Second-order thinking and 10/10/10
- Trace consequences past the first order: the obvious result, then the result of that result, then longer-term systemic effects. Many bad decisions look good only at first order.
- Run 10/10/10: how will this choice feel in 10 minutes, 10 months, and 10 years? It separates transient discomfort (a hard conversation) from durable regret (a decade in the wrong career).
### Step 7: Pre-mortem and bias sweep
Assume it is one year later and the decision failed badly. Write the story of why: warning signs, what was overlooked. Mitigate the top 2-3 risks before committing. Then check the four classic biases explicitly:
- Confirmation: name one piece of disconfirming evidence actively sought.
- Sunk cost: strike any reasoning that cites past spend; only future value counts.
- Anchoring: question whoever supplied the first number or framing.
- Overconfidence: attach an honest confidence percentage to the expected outcome.
Finish with inversion: ask "what would guarantee failure?" and confirm the chosen option avoids those things.
### Step 8: Decide, commit, and journal
Make the call by the deadline and record a decision-journal entry:
```
- DECISION JOURNAL — [FILL: date]
+ DECISION JOURNAL - [FILL: date]
Decision: [FILL: one sentence]
Options considered: [FILL]
Choice and why: [FILL: top 3 reasons]
- Expected outcome: [FILL] — confidence: [FILL]%
+ Expected outcome: [FILL] - confidence: [FILL]%
Key assumptions: [FILL: the 2-3 that, if wrong, flip the decision]
State of mind: [FILL: tired / rushed / calm / excited]
Review date: [FILL: 6-12 months out]
```
Review entries months later to separate good decisions from lucky outcomes.
## Deliverable
Produce a one-page decision brief containing: the reversibility classification, the framed question, the weighted criteria matrix with scores, expected-value math for any probabilistic payoff, the second-order/10-10-10 read, the pre-mortem top risks with mitigations, a clear recommendation, and a filled decision-journal entry.
## Do NOT
- - Do not run the full process on a two-way door — over-analysis of reversible choices is its own failure mode.
+ - Do not run the full process on a two-way door - over-analysis of reversible choices is its own failure mode.
- Do not score options before weights are locked; that is how a favorite gets rationalized.
- Do not present a weighted total under a 0.3 gap as a verdict; call it a tie and interrogate the weights.
- Do not let equity, lottery-style upside, or "best case" numbers into the comparison without explicit probabilities attached.
- Do not skip the journal entry; without it there is no way to calibrate judgment later.
## Quality bar
- The reversibility classification is stated and justified before any analysis.
- Criteria and weights were written before options were scored.
- Every uncertain payoff carries an explicit probability the user agreed to.
- The pre-mortem names at least two concrete failure modes with mitigations.
- The journal entry is filled in, including confidence and state of mind, with a review date set.