English | 简体中文
Investment Knowledge Base
A principle-driven investment framework built with Obsidian, plus the sanitized mirror of a real, running US-equity trading system. The foundation is the stock price formula:
$$P = EPS \times PE$$
- EPS (Earnings Per Share): Business fundamentals and intrinsic growth (micro perspective)
- PE (Price/Earnings): Market expectations and liquidity (macro + sentiment)
The trading side is built on a different premise: entry authorization comes from price structure only; decisions are probabilistic; exits are mechanical. Fundamentals, macro and valuation inform the prior — they never grant a trade.
Motto — Better to miss a trade than to lose your rhythm; the trader who loses rhythm ends up as the market's prey.
Structure
00_Index/ → Dashboard and navigation
10_Core_Logic/ → First principles: Davis Cycle, EPS, PE, earnings surprise
20_Macro/ → Fed policy, treasury yields, monetary history, gold
30_Micro_Analysis/ → Moats, Porter's Five Forces, financial health, tech metrics
40_Valuation/ → PE Band, DCF, relative valuation, metric selection
50_Psychology_Risk/ → Sentiment, position sizing, stop-loss, exit strategy
60_Strategy/ → Sector thesis and annual investment plan (example)
us-stocks/ → The live trading system (sanitized mirror)
├─ system-overview.md one-page map of the whole system
├─ strategy/ methodology canonical (+ options/, directives/, rules/)
├─ methods/ macro & fundamental methods (zero trading authority)
├─ research/ pre-registered backtests: what was adopted, what was rejected
├─ reviews/ anonymized real case studies
└─ knowledge/ thematic essays
polymarket/ → Archived template workspace (not maintained)
Highlights
- System Overview — division of labor (human picks names / system plans entries / system's clock owns exits / human executes), the decision stack, sizing, exit stack and daily cadence on one page
- Two entry modes only — 🔁 bottom-fishing on seller exhaustion (stop = the reversal low) and 🚀 trend confirmation via a gate-transition state machine; everything else is off the table
- Cards as state machines — every name carries zones, a binary regime cap {0, 40% NAV} and the next gate; gates are discovered where the market repeatedly made decisions, not picked
- Six-arm sizing — regime cap, gate risk budget (2.5 / 4 / 5% NAV), at-risk cap, shock budget, gross cap and buying power; the tightest arm wins and is always reported
- Mechanical exit stack, no overrides — card stop, profit ratchet, chandelier trend line, option ladder; "watch one more day" was tried three times, paid every time, and abolished
- Research verdicts — pre-registered, often out-of-sample tests; many intuitive ideas (key-level breakouts, extra bottom "evidence", tighter volume thresholds) were rejected by the data
- Post-trade reviews — real, anonymized cases; process is graded separately from outcome (wrong process + good outcome = D, the most dangerous grade)
How to Use
- Open in Obsidian for the full experience (bidirectional links, Mermaid diagrams, LaTeX)
- Start at
00_Index/Home.mdfor the knowledge framework, orus-stocks/system-overview.mdfor the trading system - Use
us-stocks/holdings/current/_EXAMPLE_.mdas a template for your own position cards - Use
us-stocks/strategy/trading-rules.md(master index) as your pre-trade reference
Anonymization
Tickers are replaced with STOCK_* placeholders, account-level figures are removed or expressed as ratios, and company-identifying catalysts are generalized. Names such as quant/... refer to code enforcement points in the author's private repository — they are kept to show where a rule is enforced ("a rule that isn't in code doesn't exist").
Philosophy
Process over outcome. Discipline compounds. Build the system, fight the randomness.
License
CC BY-SA 4.0 — share and adapt with attribution.
Disclaimer
This is a personal learning framework, not financial advice. All example positions, prices, and reviews are anonymized. Do your own research and consult qualified professionals before making investment decisions.