R — Original Text (Reading)
I have truly invested in at most five or six companies from beginning to end. I sold some, and I generally hold around three companies. Buffett's Berkshire has a market value of more than a hundred billion dollars, and he has only invested in about ten companies. I am not afraid of concentration. I am not ordinarily concentrated; I am absolutely concentrated.
Investment diversification is a false proposition. So-called investment diversification refers to wealth management; it and investing are two different games. ... Since you think it is the best and you only invest a small part, and then put the remaining money into companies you think are not good enough, that logic is simply wrong.
— Duan Yongping, Chapter 3, Section 4, "Diversify or Concentrate" / Interview with Caijing
I — Methodological Skeleton (Interpretation)
The underlying logic of concentrated investing is a simple deduction: if you truly understand a company (within your circle of competence), you should put your money into the best opportunity; if you simultaneously hold many stocks, it means you do not understand any of them well enough.
- Diversification = protection against ignorance: If you need to buy 10 stocks to diversify risk, it precisely shows that you do not understand each one well enough. True understanding brings certainty, not a need for diversification.
- Position size reflects depth of understanding: The position percentage for a stock should be determined by the depth of your understanding, not by a "risk management" formula. "The usual ratio is 0%-100%, depending on how much you yourself understand."
- The more companies you invest in, the less you earn: For holdings beyond your circle of competence, each additional one dilutes your attention to each of them and increases the probability of making mistakes.
- 20 decisions in a lifetime: If you could invest only 20 times in your life (the punch-card metaphor), you would treat every move with extreme caution. Fewer decisions, weightier decisions.
- Concentration is simple but extremely difficult: What is difficult is not concentration itself, but that "truly understanding a company" is itself extremely difficult. Understanding even one stock is very hard; ten is almost a "mission impossible."
- With little money, you especially need to focus: When the amount of capital is small, there is even less capital to diversify; you should concentrate on the best 2-3 opportunities.
- Diversification is called wealth management, not investing: Sprinkling money around in a diversified way is wealth management (keeping up with the market), not investing (finding good companies and taking heavy positions).
A1 — Application in the Book (Past Application)
Case 1: Duan Yongping's extremely concentrated holdings
- Problem: How does Duan Yongping manage his own investment portfolio?
- Use of the methodology: From beginning to end, he truly invested in at most five or six companies, and generally holds around three companies (Apple and Moutai as the core). When buying NetEase, he put in all the money he could use at the time. His Apple position was gradually increased to 60-70%.
- Conclusion: "It is so hard to find a good stock at a good price; not concentrating is a sign of not understanding."
- Result: NetEase returned about 100x, while Apple and Moutai were both long-term heavily held major winners. Munger commented on Berkshire, "Strip out 10 investments and it's a joke"—over 57 years of history, the returns from concentrated investing came almost entirely from a few decisions.
Case 2: Punch Cards and Decision Frequency
- Question: At what frequency should investors make investment decisions?
- Use of the Methodology: Duan Yongping cites Buffett's punch-card metaphor—a card with only 20 holes, punching one hole per investment, with only 20 opportunities in a lifetime. He himself reflected, "In these 4 years, Dadao has only truly acted once, which was Apple; the previous time was GE," averaging about one company every two years. In 10 years, he roughly understood fewer than 10 companies and made heavy investments in 5.
- Conclusion: "When you're sure, you have to act hard. If you half-understand, it's hard to act hard."
- Result: An extremely small number of actions brings extremely high quality per decision and concentration of returns.
A2 — Trigger Scenarios (Future Trigger) ★
In what situations will users need this skill?
- Hesitating over the number of holdings when building a portfolio — The user has a sum of money and is agonizing over "how many stocks is appropriate to buy," whether to diversify into 8-10 or concentrate into 2-3
- Already holding positions but hesitating to add because the position is too light — The user is bullish on a stock but only dares to put 5%-10% in it, wants to "balance the risk a bit," and is asking how the position should be sized
- Holding too many stocks, energy is dispersed, so each one is not researched deeply — The user holds more than 10 stocks, finds that they only "half-understand" each one, feels anxious frequently, and wants to streamline the portfolio
Language Signals (activate when these appear in the user's words)
- "I bought 10 stocks to diversify risk" / "Should I buy a few more to balance things out"
- "I only dare to put a 5% position in this stock" / "I don't dare buy too large a proportion"
- "How come the more companies I invest in, the less I earn" / "Why do I feel more anxious the more stocks I hold"
Distinction from Adjacent Skills
- Difference from
circle-of-competence(Circle of Competence decision framework): the circle of competence addresses "do I understand this stock," while this skill addresses "after understanding it, how much should I invest." The circle of competence is the premise; concentrated investing is the position-sizing corollary of the circle of competence. - Difference from
opportunity-cost(Opportunity Cost decision framework): opportunity cost addresses the comparison problem of "hold A or switch to B," while this skill addresses the quantity problem of "how many should I hold in total." But the two are logically interconnected—one reason for concentrated investing is opportunity cost: every dollar should be invested in the best opportunity.
E — Executable Steps (Execution)
Once the skill is activated, the agent should execute the following steps:
-
Diagnose the depth of understanding of current holdings
- For each holding, ask the user to answer: Can you judge this company's profitability over the next 5-10 years? If you can clearly describe its business model, competitive advantages, and risk points → understanding level meets the standard; if you can only say "I'm bullish on this industry" or "I heard it's good" → it is outside the circle of competence.
- Completion standard: The user gives a clear judgment of "understand" or "don't understand" for each holding and lists the basis for understanding.
-
Reallocate positions according to level of understanding
- For positions you "don't understand" → it is recommended to liquidate them and concentrate funds in targets you do understand. For positions you "understand" → the position ratio should reflect the depth of your understanding and confidence in it, rather than being equally diversified. Duan Yongping's standard: "0%-100%, depending on how much you yourself understand."
- Completion criteria: The user lists 2-3 companies they truly understand and gives a position allocation plan based on their level of understanding.
- Stop condition: If the user finds that they are "not sufficiently knowledgeable" about any stock, skip to Step 3 and recommend not investing for now, returning to the learning stage within the circle of competence.
-
Establish the discipline of "few decisions, heavy decisions"
- Use punch-card thinking: Assume you can only invest 20 times in your entire life—is this move worth using up one slot? If not → keep waiting. If yes → dare to bet heavily. Before each new position, first ask yourself: Is this better than the companies I already hold?
- Completion criteria: The user can state the reason why each of their current holdings is or is not "worth a punch-card slot."
B — Boundary ★
Do not use this skill in the following situations
- Capital is so large that a single target's liquidity cannot accommodate it: When your capital scale exceeds a significant proportion of the target company's daily trading volume, concentrated positions face liquidity risk, and at this point you must diversify appropriately (this is why Buffett, with large-scale capital, must also diversify).
- Asset allocation is needed rather than stock picking: If the user's need is goal-oriented financial planning such as "retirement planning" or "children's education funds," rather than active stock-picking investment, diversified allocation (stock-bond balance, index funds) is a more suitable strategy.
- Complete beginners who do not understand investing at all: For people with no foundation in investing, concentrated investing is dangerous—they easily mistake "taking a gamble" for "concentrated investing." Duan Yongping himself also said, "There are many people who never understand their whole lives, so they never touch stocks their whole lives; they are also the best value investors."
Failure modes the author warns about in the book
- Mistaking gambling for concentrated investing: Duan Yongping clearly distinguishes—investment concentration is because you "understand it." If you concentrate without understanding, that is called gambling. "If you didn't understand it, what kind of move are you making?"
- Betting heavily when you half-understand: "It is hard to act ruthlessly when you half-understand"—only true understanding allows heavy positions; for people who think they understand but actually do not, the result of concentrated holdings is disastrous.
- Pseudo-concentration with frequent position switching: Concentrating in this one today and that one tomorrow is essentially speculation, not concentrated investing.
The author's blind spots / limitations of the era
- Differences in capital size: As a large investor, Duan Yongping's concentration strategy is built on the basis of "taking many years to understand." People with small capital may face a different opportunity structure—good opportunities are fewer, but the returns from concentrated bets are larger, while the absolute loss risk they bear is also more controllable. Conversely, however, people with small capital are more likely to suffer devastating blows from concentrated investing (with no other source of income to hedge).
- Survivorship bias: Successful cases of concentrated investing (Duan Yongping, Buffett, Munger) are widely spread, but people who fail at concentrated investing do not write books. Duan Yongping himself also admitted that NetEase was an opportunity "possibly never to be encountered again in this lifetime."
- The blurriness of the boundaries of the circle of competence: The judgment of "do you really understand or not" is itself extremely difficult—many people think they understand but actually do not. The premise of concentrated investing is the accuracy of the circle-of-competence judgment, and this premise itself carries uncertainty.
Adjacent methodologies that are easily confused
- Index investing: Buffett himself recommends that most people buy an S&P 500 index fund—this seems to contradict concentrated investing. The difference lies in this: index investing acknowledges "I cannot pick good companies," and is rational diversification; the premise of concentrated investing is "I can pick good companies." If your circle of competence is not sufficient to pick good companies, concentrated investing is gambling.
- All-in style gambling: Concentrated investing is not going all in. The difference lies in the premise: concentrated investing is a rational heavy position taken after the three conditions of circle of competence + good business + reasonable price are simultaneously met; gambling is going all in without understanding anything.
Related skills (to be filled in stage 3)
- depends-on: [capability-circle, good-business]
- contrasts-with: []
- composes-with: [dcf-valuation]
Audit information
- Verification passed: V1 ✓ / V2 ✓ / V3 ✓
- Test pass rate: Pending stage 4 (see test-prompts.json for details)
- Distillation time: 2026-04-16