Wang Kai RSI AI Investment Strategy and the "Good Asset, Good Price, Good Volatility" Framework is Philosophy Layer(Why, principles of the RSI AI framework); this wiki is Practical Layer(How to do it: 5 specific cases + deepening of value investment philosophy + Prompt template).
Timeframe: 2026-01 ~ 2026-02, 10 deliverables from Wang Kai's deep application phase.
I. Comprehensive Prompt Template for Corporate Intrinsic Value Analysis (file_041 → docx 16KB)
Wang Kai instructs AI to first quantify "good assets" before deciding capital allocation, via this prompt template.
Role Setting
You are a senior equity research analyst with cross-industry, cross-market deep research capabilities. You are proficient in multiple valuation methodologies, able to select the most suitable analysis framework based on corporate characteristics, and draw reliable conclusions through cross-validation of multiple methods.
8-Part Structure
- Corporate Classification and Method Matching — 4 corporate types: Mature SaaS / High-growth platform / Hardware + service ecosystem / Diversified conglomerate
- Business Quality Analysis — Business model / Competitive position (Porter's Five Forces) / Growth drivers / Management and governance
- Financial Analysis — General financial metrics / Industry-specific metrics / Cash flow quality / Balance sheet
- Detailed Valuation Methods — Five methods: DCF / Relative valuation / DDM / Shareholder return model / Reverse DCF
- Valuation Synthesis and Investment Decision — Method weight allocation + price range
- Risk Assessment — 4 major risk categories: Business / Regulatory / Financial / Macro + scenario analysis
- Investment Conclusion and Recommendations — Bull / Bear / Key variables
- Ongoing Tracking — Key indicator monitoring table + event calendar
Application Conventions
- Input: Stock ticker + analysis purpose (full valuation / buy-sell decision / deep research / comparative analysis) + special focus areas
- Output: Each section filled independently,cross-validated with multiple methods, not relying on any single method
II. BKNG Case (Booking Holdings) — RSI Trigger + Full Valuation Review
Trigger Signal (2026-02-19, file_047)
RSI-7 = 20.51 (极度超卖 <30)
RSI-14 = 23.15 (超卖)
RSI-21 = 26.47 (接近超卖)
Williams %R = -100 (绝对地板)
ADX = 38.01 (强下降趋势)
价格 -8.36% 当日 / -33% 自高点
新 52 周低点
5-dimension scoring:
- Macro/Liquidity: 0.60 (neutral to positive)
- Market Sentiment: 0.38 (insiders sold 272, bought 0 → extremely bearish)
- Industry Health: 0.38 (AI disrupting OTA narrative)
- Company Quality: 0.62 (contradiction between Q4 revenue beat and EPS miss)
- Tech/Momentum: 0.35 (high probability of oversold rebound)
- TOTAL: 0.48, Risk Level 3
Position recommendation: 1 share ($3,913 = 3.9% of a $100K portfolio), split 50%+50% in batches
Deep Valuation (file_048)
Diversified cross-validation of valuation methods:
- DCF (35% weight): $5,680 base case
- Relative valuation (25% weight): $5,750 base case
- Shareholder return model (20% weight): $5,500 base case
- Reverse DCF (10% verification): implies extreme pessimism
- DDM (10% floor): $1,060
Weighted base case $5,660, current price $4,012 → +41% upside
8-part structure expansion(full application of file_041 template):
- Corporate classification: Mature high-growth platform + consumer goods attributes
- Business quality score: 51/60 (excellent)
- Financial quality: FCF $9.1B / EBITDA Margin 36.9% / Net debt ≈ 0
- Industry: GMV $186.1B / 1.2 billion room nights / merchant model share 63%→68%
- Moat rating:Wide Moat(brand + network effects + economies of scale + data flywheel)
- Risks: AI disrupting OTA (Google Travel), EU DMA, macro cycles
Key judgment: Operation recommendation table
| Stock Price | Operation | Position |
|---|---|---|
| < $3,800 | Aggressive Buy | 8% |
| $3,800-4,200 | Gradual Position Building | 6% |
| $4,200-5,200 | Hold | Maintain |
| $5,200-6,000 | Partial Profit Taking | 4% |
| > $6,000 | Reduce Position | 2% |
III. AMZN Case (Amazon, file_042)
Trigger: 2026-02-05 close at $222.69
RSI:
- RSI-7 = 24.52 (oversold)
- RSI-14 = 35.96
Key Data:
- Q4 EPS $1.95 vs $1.97 expected (slight miss)
- Q4 revenue $213.4 billion vs $211.3 billion expected (beat)
- AWS revenue $35.6 billion (+24% YoY)
- 2026 capex guidance $200 billion vs $146 billion expected — massive computing power investment
- Q1 operating profit guidance $16.5–21.5 billion vs $22.0 billion expected
Bear: Staggering capex + weak Q1 guidance + concerns that AWS growth is starting to slow Bull: AWS remains the flagship + long-term AI dividends
(Detailed filling of the 8-part template omitted; RSI framework + 5-dimension scoring are universally applied)
IV. QQQ Case (Nasdaq 100 ETF, file_043)
Trigger:2026-02-27,$607.29
- TTM P/E ~33.4x vs 5-year average ~28x (significantly expensive)
- Forward P/E ~27x(considering AI-driven 2026 EPS +15–20%) is more reasonable
- 52-week range: $402.39–$637.01
Judgment:Reasonably to slightly expensive, suitable for long-term holding but note near-term pullback risk
Valuation Methods:
- Weighted P/E (forward): $550–670 / 30% weight
- Earnings growth discount: $520–700 / 25% weight
- Relative valuation (vs SPY/DIA): $480–650 / 20% weight
- Reverse DCF: (not expanded)
V. SPY Case (S&P 500 ETF, file_044)
Trigger:2026-03-03,$679.19
- 52-week range $481.80–$697.84
- TTM P/E 27.74 / Forward P/E 21.6 / Shiller CAPE 39.8 (second-highest peak in history, only behind the 2000 dot-com bubble)
- Dividend yield 1.15%, AUM $698B, VIX 22.40
Key Judgments:
- Forward P/E 21.6 > 5-year average 20.0 > 10-year average 18.7
- CAPE 39.8 → second-highest in history
- CY2026 expected EPS +14.9% (strong earnings)
- Rising macro uncertainty (VIX 22.40 / trade policy / sticky inflation)
Conclusion:Risk-adjusted return attractiveness declines
Valuation:
- DDM: $580 bear / $630 base / $700 bull, 25% weight
- Relative valuation: $590 / $650 / $720, 30% weight
- Shareholder return model: $570 / $620 / $690, 20% weight
VI. SONY Case (file_040, xlsx)
Trigger:2026-01-29
- Stock price $22.28
- Market cap $133.1 billion
- TTM PE 17.14x
- 52-week $20.42–$30.34
- Beta 0.801
RSI:
- RSI-7 = 9.43 (extremely oversold)
- RSI-14 = 16.78 (severely oversold)
- RSI-21 = 23.23 (oversold)
- Williams %R = -95.31 (oversold)
- ADX 43.28 (trend confirmed)
Valuation Methods(full 5-method weighting):
- DCF discounted cash flow: $24.50 bear / $28.75 base / $35.20 bull, 30% weight
- SOTP sum-of-the-parts valuation: $26.80 / $31.50 / $38.00, 25% weight
- Relative valuation method: $25.00 / $29.50 / $34.00, 20% weight
- Dividend discount model: $23.00 / $26.00 / $30.00, 10% weight
- Shareholder return model: $24.00 / $27.50 / $32.00, 15% weight
Weighted base case $29.15 vs current $22.28 → +30.8% upside
Investment Rating:Buy(base case $29.15) Risk Level: Medium
VII. Deepening of Value Investment Philosophy in the AI Era (file_037, refined from 120+ rounds of dialogue across 6 dimensions)
Refined through 120+ rounds of internal dialogue among 6 deep-research Agents.An order of magnitude deeper than a single Skill call.
First Principles (eternal and unchanging)
| Classic Principles | How to uphold them in the AI era |
|---|---|
| Buying stocks is buying pieces of a company | Uphold even more firmly — do not get swayed by AI-driven trading rhythms |
| Margin of safety | Uphold even more firmly — because uncertainty is greater |
| Circle of competence | Uphold even more firmly — beware of false expansion of the circle of competence |
| Long-term perspective | Uphold even more firmly — navigate more violent volatility |
| Integrity / Honesty | Uphold even more firmly — AI amplifies the destructive power of dishonesty |
| Contrarian thinking | Even more valuable — consensus forms more easily to excess |
New types of moats in the AI era (ranked by reliability)
First tier: Slow-changing layers that AI can barely touch
- Trust capital (reputation built over decades of consistent words and deeds)
- Regulatory + physical asset portfolios (licenses + scarce assets)
- Organizational execution (culture + talent + processes)
Second tier: Enhanceable by AI but not easily replaced 4. Deep customer embedding 5. Personality-based brands (trust built on real personas) 6. Interpersonal network effects
Third tier: Evaluate with caution 7. Unique data flywheels 8. AI ecosystem effects
Fourth tier: Potentially weakened 9. Functional network effects 10. Pure economies of scale 11. Codifiable knowledge barriers
5 Major Dangerous Traps
| Trap | Danger |
|---|---|
| AI endorsement psychology | Because AI said it, you feel more confident |
| Complexity trap | AI can perform complex analysis → makes people think "the more complex the analysis, the better" |
| Information overload | Too much information leads to analysis paralysis |
| Groupthink | Everyone uses similar AI → similar conclusions |
| Circle of competence illusion | AI makes you think "I understand", when in reality only "AI understands" |
Choose one of three investment strategies
| Strategy | Suitable for | Specific approach |
|---|---|---|
| Avoidance strategy | Most investors | Invest in traditional industry leaders that will not be disrupted by AI |
| Beneficiary strategy | Those with some AI understanding | Invest in traditional companies that can leverage AI |
| Direct investment strategy | Deep AI understanding | Directly invest in AI companies (requires extremely large margin of safety) |
Recommendation: For most value investors = a combination of "avoidance + beneficiary"
VIII. AI-Native Investment Paradigm (file_038, 6 major principles)
Different from file_037, this document is"breaking away from traditional frameworks" to explore new AI-native paradigms.
5 Major Principles
| Principle | Core |
|---|---|
| Questions are more valuable than answers | AI lowers the cost of finding answers → the ability to ask the right questions becomes scarcer |
| Time is the ultimate moat | AI accelerates everything → "slowness" becomes a scarce resource |
| Human nature is the unchanging anchor | Technology changes, human nature does not |
| Metacognition is the core capability | Knowing when to listen to AI and when to trust yourself |
| Individual + AI is the new investment subject | A single person can possess capabilities previously only available to institutions |
4 Core Paradoxes
| Paradox | Description |
|---|---|
| 1 | AI makes information more transparent, but judgment harder |
| 2 | AI makes analysis easier, but understanding harder |
| 3 | AI makes speed faster, but slowness becomes more valuable |
| 4 | AI makes complexity easier, but simplicity becomes scarcer |
New types of scarce resources (5 kinds)
- Relational capital: Real interpersonal trust cannot be replicated by AI
- Courage capital: The ability to act amid fear is always scarce
- Time capital: The ability to wait patiently becomes scarcer
- Creativity capital: Defining new dimensions rather than optimizing within known dimensions
- Physical world capital: AI is omnipotent in the digital world; the physical world requires real existence
One-sentence summary
Investing in the AI era is not about using AI to replace human judgment, but about using AI to amplify people's unique advantages — building lasting competitive edges in areas that AI cannot reach, such as human nature, time, trust, and courage.
IX. Quick Framework for Core Insights (file_039)
Simplified version, changes across 8 major dimensions:
| Dimension | Old Paradigm | New Paradigm |
|---|---|---|
| Decision-making method | Humans make judgments alone | Human-machine collaboration: humans set rules, AI executes judgments |
| Source of advantage | Information gap, speed | Disciplined execution, cognitive iteration, system capability |
| Risk management | Vague intuition | Quantified, traceable, consistent |
| Emotion handling | Overcome by willpower | Outsourced to AI systems |
| Strategy execution | Frequently deviates from original intent | Consistent execution, reduced human intervention |
| Learning method | Rely on experience and memory | Systematic recording, data analysis, continuous review |
| Time allocation | Large amount of time spent on research and analysis | Focus on rule design and final decision-making |
| Focus of competition | Who is smarter | Whose system is better |
One-sentence essence
Investing in the AI era is not "using AI to predict the market" but"using AI to conquer yourself". Markets are becoming increasingly efficient, yet most people cannot even conquer themselves.
X. Common Methodology of the Five Cases
Application steps for the 5 cases (BKNG / AMZN / QQQ / SPY / SONY):
1. RSI 触发器扫描(主战场美股,每日盘后)
→ RSI-7 < 30 + Williams %R < -80 + ADX > 25
→ 触发"深度分析"流程
2. 应用 5 维度评分框架(file_041 内化的 Skill)
- Macro/Liquidity (动态权重)
- Market Sentiment (内部人交易 / 机构 / 散户)
- Industry Health
- Company Quality (Q earnings + 财务三表)
- Tech/Momentum (RSI + ADX + 均线)
3. 5 种估值方法交叉验证(file_041 → file_048 完整模板)
- DCF / 相对估值 / DDM / 股东回报模型 / 逆向 DCF
4. 加权基准 + 上涨空间计算 → 风险等级判定
5. 操作建议表 (按价格区间分级建仓)
6. 关键日期 + 监控指标表 → 持续跟踪
Key constraints:
- Position per operation ≤ 5% (capital management)
- 100% cash floor (protection against zero-risk)
- Phased entry (50% + 50%)
- Stop-loss rule 1.1× ATR
See also
- Wang Kai RSI AI Investment Strategy and the "Good Asset, Good Price, Good Volatility" Framework — Principle layer of Wang Kai's investment strategy (this wiki is the practical layer)
- Wang Kai's AI-Era Decision-Making System — AI as a ruler to quantify the source of vague risks
- Wang Kai Collection / Close Reading of 2024–2026 Cutting-Edge AI Papers and Reports — Theoretical support including the Aschenbrenner timeline, Hallucination paper, etc.
- Wang Kai IFS (Internal Family Systems) AI-Assisted Practice + Claude Skill Definition — Emotion management in investing (IFS inner work)