Prompt de Análisis de Inversión de Howard Marks

First-level thinking is simple, superficial: "This is a good company, let's buy the stock."

Second-level thinking is complex, convoluted:
- "This is a good company, but everyone thinks so, so stock is overvalued"
- "Market expects 20% growth, but I think only 15%, so should sell"
- "This company is terrible, but market overly pessimistic, reality not that bad"

**Core:**
- Don't ask "Is this a good company?", ask "Is market pricing it correctly?"
- Superior investing not buying good assets, but **buying well**
- Consider probability distributions, not just single outcomes

Marks' market cycle pendulum:

**Late bull market signals:**
- "This time is different" becomes popular phrase
- Everyone talking about stocks
- Novices flooding into market
- Valuations at historical highs, but "rationalization" explanations prevail
- Leverage widely used

**Late bear market signals:**
- Pessimistic rhetoric of "stock market is dead"
- Quality assets ignored
- Panic selling
- Media full of doomsday predictions
- Almost no one dares to buy

**Key indicators:**
- Investor sentiment (greed vs fear)
- Credit environment (loose vs tight)
- Risk appetite (high vs low)

**Traditional definition (wrong):** Volatility
- Academia measures risk with standard deviation
- But Marks thinks this is wrong

**Marks' definition:** Possibility of permanent loss
- Risk is not stock price fluctuation
- Risk is **permanent loss of capital** (company bankruptcy, never recovers)

**Counter-intuitive views:**
- 30% stock price drop not increased risk, possibly increased opportunity
- 50% stock price rise not decreased risk, possibly increased risk
- "Safe" government bonds also have risk during inflation

**Marks quotes:**
"Risk is greatest when people think it doesn't exist."
"Risk is lowest when people think it's everywhere."

⚠️ **Cannot precisely time, but can avoid big mistakes.**

**Cannot do:**
- Precisely predict tops and bottoms
- Tell you "buy next week" "sell next month"
- Always buy at lowest point

**Can do:**
- Raise vigilance during bull market mania (reduce buying, increase selling)
- Aggressively attack during bear market panic (increase buying, reduce selling)
- Avoid being fully invested at market peak, fully cash at market bottom

**Marks' strategy:**
- Oaktree held massive cash 2005-2007 (though missed some gains)
- Aggressively bought distressed assets 2008-2009 crisis
- Result: Stunning returns 2009-2012

**Key: Not prediction, but positioning** (knowing where we are in cycle)

✅ **3 practical methods:**

**1. Build contrarian investment checklist:**
- Bull market mania: Gradually reduce positions, stop chasing, increase cash
- Bear market panic: Buy in batches, buy quality assets, reduce cash

**2. Watch market sentiment indicators:**
- VIX fear index (high fear = buy signal)
- Margin debt (leverage too high = danger signal)
- New fund issuance (spikes at bull market end)

**3. Ask yourself second-level questions:**
- Don't ask "Will this stock rise?"
- Ask "What is market consensus? How does my view differ from consensus? If I'm right, how will market react?"

⚠️ **Challenge: Emotional control**
- Hard to stay calm during market mania
- Buying during market panic requires enormous courage

✅ **Most suitable scenarios:**
- **Market extremes**: 2008 financial crisis, 2020 pandemic crash, bull market bubbles
- **Asset allocation**: Deciding stock/bond/cash ratios
- **Distressed investing**: Buying quality assets oversold
- **Risk management**: Controlling positions, avoiding full investment or full cash

❌ **Not suitable scenarios:**
- Selecting specific stocks (Marks mainly does asset allocation and distressed debt)
- Short-term trading (cycle judgment is medium-long term)
- Growth stock investing (Marks leans toward value investing)

**Combination advice:**
- Use Marks' method to control overall position sizing
- Use Buffett's method to select specific stocks
- Use Munger's mental models to avoid mistakes

**3-stage strategy in bull market:**

**Early stage (few participants):**
- Aggressively buy quality assets
- Moderate use of leverage
- Position can be 70-80%

**Middle stage (many participants):**
- Continue holding, but stop adding
- Reduce leverage, lower risk
- Control position at 50-60%

**Late stage (everyone euphoric):**
- Gradually reduce positions, lock in profits
- Increase cash ratio to 40-50%
- Reject FOMO (fear of missing out) temptation

**Marks' 2007 decision:**
- Refused to participate in subprime mortgage mania
- Held massive cash (though looked "stupid")
- Prepared for 2008-2009 opportunities

**Result:** Oaktree became one of crisis's biggest winners.

⚠️ Directional judgment has reference value, precise positioning impossible.

What AI can do:
✅ Synthesize multiple indicators to judge roughly "first half" or "second half" of cycle
✅ Identify extreme signals

What AI cannot do:
❌ Precisely determine where "top" or "bottom" is
❌ Predict when turning point will appear

✅ Correct usage:
1. Treat AI's judgment as "approximate location", not precise coordinates
2. Focus on objective indicators AI cites (CAPE, VIX, margin debt), verify yourself
3. Marks' method is not prediction, but "sensing temperature"
4. Adjust strategy based on temperature: more conservative when hot, more aggressive when cold

⚠️ Key is checking whether AI truly thinks at "second level", not just applying templates.

True second-level thinking needs to answer:
1. What is market consensus?
2. Where does my view differ?
3. If I'm right, what happens?
4. If I'm wrong, what happens?

⚠️ Common "fake second-level thinking":
- AI just says "stock might go up or down" — meaningless
- AI gives contrarian view but without sufficient supporting arguments
- AI's "second level" has actually become the new consensus

✅ Verification method:
- Read latest research to confirm AI's "market consensus" is accurate
- Verify AI's independent judgment with data
- Ask: If I posted AI's analysis on investment forum, would anyone disagree? If not, probably not true second-level thinking

✅ Marks-style action framework:

1️⃣ Determine current "market temperature"
- Combine AI analysis with your own observations
- Rate current market: 1 (extreme panic) to 10 (extreme greed)

2️⃣ Adjust positions based on temperature
- Temperature 1-3 (panic): Actively buy, position up to 70-80%
- Temperature 4-6 (neutral): Maintain normal position 50-60%
- Temperature 7-9 (greed): Gradually reduce, cash to 30-50%

3️⃣ Execution discipline
- Don't be influenced by FOMO
- Don't try to precisely catch the bottom
- Operate in batches

4️⃣ Continuous tracking
- Re-evaluate market temperature monthly
- Record judgments and actual results
- Optimize temperature assessment from mistakes