Milestone map
Milestone map
3 milestones
Audit Your Real Decision Failures
1–2 weeks (2–3 hrs total)
Choose one domain where your decisions have mattered (work, finances, a project, health). Review the past 12 months and identify at least 5 decisions where the outcome was worse than you expected. For each, document four things: the decision made, what you believed at the time, what actually happened, and whether the gap arose from a process failure, bad luck, or inadequate information. At the end write a 150-word analysis naming your one most common process failure pattern.
Proof required
Submit your decision audit: a structured document containing at least 5 real past decisions, each with four labelled fields (decision made / belief at the time / actual outcome / root cause: process | luck | information). At least 2 must be classified as process failures with a specific explanation of what process was missing. Append your 150-word signature analysis naming one recurring pattern.
What gets checked
- 5 real decisions with all four fields completed — generic descriptions such as 'made a bad hire' without the specific reasoning at the time do not count
- At least 2 root-cause attributions of 'process failure' with a named specific gap (e.g. 'skipped reference checks', 'no pre-mortem run')
- Signature analysis names one pattern with evidence from at least 2 of the 5 entries — a general observation not tied to the entries is insufficient
Resources
Foundationstart here
Depthgo deeper
What a verifier looks for
- Ask the submitter to describe one of the process-failure entries in detail — what specifically did they skip and why at the time?
- Check that root-cause labels are applied with reasoning, not just assigned to reach the '2 process failures' minimum
- Confirm the signature pattern named in the analysis matches 2+ specific entries rather than being a general observation
- A list of decisions classified only as 'bad luck' with no process failures named is a red flag — almost every decision audit surfaces at least some process issues
- Attributing every bad outcome to bad luck or external factors to avoid examining one's own process
- Selecting low-stakes decisions where errors carry no real cost — the audit should cover decisions with material consequences
- Writing the document retrospectively from memory without checking any contemporaneous records (emails, notes, calendar)
Run a Decision Framework for 8 Weeks
8 weeks (1–2 hrs/week)
Choose one structured decision framework — pre-mortem, second-order consequences mapping, confidence calibration (assign a percentage probability before deciding), devil's advocate, or decision journaling as defined by Annie Duke — and apply it consistently to every significant decision you face for 8 consecutive weeks. Document each application contemporaneously: what the framework surfaced that you had not initially considered, whether it changed your inclination, and what you decided. Aim for at least 1–2 applications per week.
Proof required
Submit your 8-week decision journal: at least 12 documented decisions, each with four fields (decision faced / what the framework surfaced / whether inclination changed — yes/no + how / decision made). Entries must be dated and spread across the 8-week period. The same framework must appear across all 12 entries.
What gets checked
- 12 or more documented decisions with all four fields — a summary of the 8 weeks written retrospectively does not qualify
- Framework visibly changes inclination on at least 4 of the 12 entries with a specific explanation of how, not just 'yes it helped'
- Dates on entries are spread across the 8 weeks — entries clustered in the final week indicate retrospective writing
Resources
Foundationstart here
Depthgo deeper
What a verifier looks for
- Confirm the same framework appears across all 12 entries — switching frameworks between entries is a failure mode, not a sign of sophistication
- Look for at least 4 entries where 'inclination changed' is 'yes' with a specific explanation — if all are 'no', the framework was applied without honest engagement
- Check that entry dates span the 8 weeks — entries written in the last 2 days of the period indicate the journal was reconstructed from memory
- Ask the submitter to name a decision that was harder to make because of the framework, not easier — this reveals genuine engagement
- Switching frameworks mid-journal when one feels uncomfortable — the proof of value is in repeated application of a single framework until it becomes natural, not finding the 'right' one
- Documenting trivial decisions (what to have for lunch) to reach the minimum count rather than decisions with real stakes or uncertainty
- The 'inclination changed' field answered 'no' across all 12 entries — this suggests the framework was applied as a ritual rather than engaged with honestly
Present Real Cases Under Adversarial Review
2–3 weeks (4–6 hrs prep + 1 hr session)
Prepare a 20-minute presentation for a qualified reviewer featuring 3 real decisions from your M2 journal: one where the framework visibly improved the outcome, one where you made a mistake despite using the framework, and one where the framework surfaced a risk you nearly ignored. The reviewer's role is to ask adversarial questions: identifying flaws in your reasoning, presenting alternative interpretations, and introducing a novel scenario you have not seen to test whether the framework is genuinely internalised or only applied mechanically when prompted.
Proof required
Submit your 3-case written brief (minimum 800 words total, one section per case) plus the Q&A record from the review session: the reviewer's questions and your responses (minimum 5 distinct questions documented), and the reviewer's written sign-off naming their role, their minimum qualification, and confirming this was a live 20-minute session with adversarial questioning.
What gets checked
- All 3 case types present — one success, one framework-despite-failure, one near-miss — not 3 successes or 3 failures
- Q&A record documents actual questions asked (not paraphrased weeks later) and your actual responses, not a clean narrative
- Reviewer sign-off includes their name, their relevant background (current or recent decision-making role with ≥5 years), and confirmation the session included adversarial challenges to your reasoning
Resources
Foundationstart here
Depthgo deeper
What a verifier looks for
- Confirm all 3 case types are present and that the failure case involves a real mistake made despite the framework — not a hypothetical or a near-miss relabelled
- Ask the submitter about a decision not in their brief: walk me through how you would apply your framework to X (novel scenario). Assess whether reasoning is genuinely internalised or rehearsed
- Ask the submitter to identify the weakest part of their own reasoning in one of the cases — this tests metacognitive awareness, the deepest level of decision improvement
- Reviewer minimum qualification: 5+ years making consequential decisions (senior manager, founder, investor, experienced consultant) — a peer without this track record and experience of real decision consequences does not qualify
- Presenting only successful applications — the failure case and the near-miss are the most revealing and must be real
- A reviewer who validates every answer without pushing back — a genuine adversarial reviewer should identify at least one gap in the submitter's reasoning
- Sign-off written after the fact without the Q&A record to support it — the record is the proof of the session, not the sign-off alone