Skill in 《What I Learned About Investing from Darwin》

Skill.md

--------|------------| | ROCE trend | ✅ Multi-year median ~55%; recent dip to 50% — still exceptional | | Leverage | ✅ Zero debt, strong free cash flow generation | | Customer concentration | ✅ Millions of retail consumers; no single customer >5% revenue | | Supplier concentration | ✅ Diversified fabric and yarn suppliers | | Industry pace | ✅ Innerwear evolves slowly; not disrupted by technology | | Management stability | ✅ Long-tenured management team, founder-mentality culture | | Competitive position | ✅ Market share grew from 66% to 70% among listed peers during COVID | Robustness score: 7/7

      **Step 4: Signal Analysis**
      - **Honest signals (trust):** 10-year ROCE consistently above 40%; zero leverage; market share gains during competitive stress; dividend growth track record
      - **Dishonest signals (ignore):** FY24 single-year revenue dip; analyst downgrades on near-term volume softness; management conference call commentary

      **Step 5: GKPI Verdict**
      → **Buy on Dip**

      A 20% price decline in a business with 50%+ ROCE, zero debt, 7/7 robustness, and a multi-decade track record is a textbook punctuation event. Page Industries is precisely the kind of "boring" consumer franchise the Darwin framework is designed for. Size the position according to conviction.

  - user: "If Page's ROCE falls to 35% next year, should I sell?"
  - ai: |
      No — a single year at 35% ROCE does not trigger a sell under the Darwin framework.

      35% ROCE is still well above cost of capital (~10–12%). One year of margin compression is business noise, not business damage. Apply the honest signal test: Is there a 3+ year trend of market share loss? Has ROCE been systematically declining with no structural recovery? If yes to both, that warrants a closer look. A single-year dip does not.

      **The three sell triggers remain unchanged:**
      1. Governance deterioration — evidence of fraud or minority shareholder abuse?
      2. Egregious capital misallocation — large debt-funded acquisition in an unrelated business?
      3. Irreparable business damage — is Jockey losing structural ground to competitors?

      If none of these apply: **hold, ignore the quarterly noise, and consider adding if price has fallen further.**

      The asymmetry favors patience: maximum loss is capped at your invested amount, but if Page returns to 55%+ ROCE, the upside is uncapped.

• thread:

  • user: "A friend recommended a retail chain that's 'turning around.' New CEO, cost-cutting program, revenue starting to recover. ROCE is currently 6% but the story sounds compelling. Should I look at it?"

  • ai: | Automatic rejection — stop at Step 1.

    This triggers the single most important rejection criterion in the Darwin framework:

    Turnaround stories are automatic disqualifiers. "A bad business with a brilliant manager remains a bad business." — Prasad's core principle, drawn from Buffett.

    Why the Darwin framework rejects turnarounds categorically: The error math is decisive. Good businesses are rare (~25% of listed companies). Turnarounds require a business in the bottom quartile to cross to the top — an event that almost never happens durably.

    Reducing Type I errors (investing in bad businesses) improves portfolio performance ~5x more than improving your ability to identify good ones. Turnarounds are exactly the Type I error the framework avoids.

    "The compelling story" is a dishonest signal:

    • New CEO narrative: zero cost to produce
    • "Revenue starting to recover": one quarter of data, not a trend
    • 6% ROCE: below cost of capital — the historical record says this business has been destroying value

    What to do instead: Set a 5-year watch. If this company delivers consistent ROCE above 20% for 3+ consecutive years, zero debt, and stable market share — the Darwin framework will evaluate it as a new business. Until then, the story is irrelevant.

    The one thing the Darwin framework asks: Don't let a good narrative override a bad record.

related_skills:

• name: "The Intelligent Investor" description: "Graham's value investing and margin of safety" url: "/en/book/intelligent-investor-graham"
• name: "The Most Important Thing" description: "Second-level thinking and risk control framework" url: "/en/book/most-important-thing-in-investing-howard-marks"
• name: "The Outsiders" description: "CEO capital allocation quality evaluation framework" url: "/en/book/outsiders-ceo-capital-allocation-thorndike"
• name: "Poor Charlie's Almanack" description: "Multidisciplinary mental models and decision-making" url: "/en/book/munger-mental-models-worldly-wisdom"
• name: "Common Stocks and Uncommon Profits" description: "15-point growth stock evaluation framework" url: "/en/book/common-stocks-uncommon-profits-stock-pickup-phisher"
• name: "Understanding Financial Statements" description: "Three-statement framework for company analysis" url: "/en/book/financial-statements-analysis-xiao-xing"
• name: "The Clash of the Cultures" description: "Index vs active fund investment culture debate" url: "/en/book/clash-cultures-investment-speculation-bogle"
• name: "Where Are the Customers' Yachts?" description: "Identifying conflicts of interest in financial advice" url: "/en/book/customers-yachts-wall-street-schwed"
• name: "Business Adventures" description: "Corporate failure patterns and human nature in business" url: "/en/book/business-adventures-analysis-brooks"
• name: "Essays in Persuasion" description: "Keynesian macroeconomic policy analysis framework" url: "/en/book/essays-in-persuasion-keynes"

Investing From Darwin: Evolutionary Principles for Permanent Ownership

This skill encodes the investment framework from Pulak Prasad's What I Learned About Investing from Darwin (Columbia University Press, 2023). Prasad is the founder of Nalanda Capital, a fund managing $5B+ in listed Indian equities since 2007 with annualized rupee returns above 20% after fees. His philosophy in ten words: we want to be permanent owners of high-quality businesses.

Core Framework

Pillar I: Avoid Big Risks — Be a Great Rejector

Darwin's most counterintuitive insight, applied to investing: the most evolutionarily successful species are those that minimize Type I errors (errors of commission — self-harm) even at the cost of more Type II errors (errors of omission — missed opportunities).

The math is stark: reducing Type I error rates by 10 percentage points improves portfolio performance by ~16 percentage points. Reducing Type II error rates by the same amount improves performance by only ~3 percentage points. The skill of rejection is worth 5x the skill of discovery.

Automatic rejection triggers: governance red flags, turnaround stories, high leverage, fast-changing industries, acquisition-hungry management, concentrated customer/supplier bases.

Pillar II: Buy High Quality at a Fair Price

ROCE as the single filter: Inspired by Dmitri Belyaev's Siberian fox experiment — select one trait, get correlated benefits for free. Historical ROCE ≥20% sustained over 10+ years cascades into: genuine competitive advantage, pricing power, management discipline, and durable economics. ROCE beats management quality, revenue growth, net margin, ROE, and DCF as screening criteria.

Multi-level robustness: Organisms survive not by strength but by robustness at multiple levels. Businesses must be assessed across 7 dimensions: ROCE trend, leverage, customer/supplier concentration, industry pace, management stability, and competitive position.

Honest vs. dishonest signals: Zahavi's handicap principle — signals costly to fake are honest. Long-term ROCE, market share trends, free cash flow generation, and balance sheet quality are honest signals. Press releases, management impressions, analyst targets, and forward guidance are dishonest signals.

Pillar III: Don't Be Lazy — Be Very Lazy (GKPI)

Gould and Eldredge's punctuated equilibrium: species remain in stable stasis for millions of years, punctuated by brief rapid change. Great businesses are similarly stable. Investment windows — when quality companies become temporarily mispriced — are punctuation events. They're rare (1–2% of holding period) but significant. When they open: buy a lot.

After buying: Good Karma, Patience, Inertia. Nalanda has sold only 10 businesses in 15+ years (one every 1.5 years average). They never sell on valuation. The only three sell triggers: governance deterioration, egregious capital misallocation, or irreparable business damage.

Supported Query Types

• Five-step Darwin assessment for a specific stock (Reject → ROCE → Robustness → Signals → GKPI verdict)
• Classify whether a piece of information is an honest or dishonest signal
• Evaluate whether to sell an existing holding
• Assess whether an industry meets the "slow-changing" investability standard
• Understand how to use historical financials instead of DCF projections

How to Use

1. Download or clone this skill folder
2. In Claude.ai, go to Settings → Skills and upload the folder
3. Describe the company you want to analyze — the skill will run the five-step evaluation automatically

Limitations

This skill provides a historical-data-based analytical framework. It does not provide price targets, DCF valuations, or specific buy/sell recommendations. Final investment decisions remain the user's responsibility. The framework is best suited for slow-changing industries (consumer goods, industrials, financials) and has limited applicability to fast-changing sectors (technology, biotech).