Technical Analysis Pedagogy

This document describes how to learn Ghost’s technical analysis system. It provides a structured learning path, prerequisites, exercises, and common pitfalls to avoid.

Learning Philosophy

Principles

1. Concepts before tools. Understand what you’re measuring before learning how to measure it.
2. Simple before complex. Master single-agent analysis before multi-agent synthesis.
3. Read before interpret. Know the data before deriving meaning from it.
4. Practice with real charts. Theory only becomes skill through application.

The Knowledge Framework

Ghost’s knowledge is organized in seven layers:

FOUNDATIONAL
├── ONTOLOGY        → What are the concepts?
├── EPISTEMOLOGY    → How does Ghost know? What are the limits?
└── ETHICS          → What must Ghost do and never do?

PER-AGENT
├── METHODOLOGY     → How do I identify them?
└── HERMENEUTICS    → How do I interpret them?

PRACTICE
├── SYNTHESIS       → How do I combine them? (Director prompts)
├── PRAXIS          → How does knowledge become action?
└── INTERROGATION   → How do we evaluate prompt signal density?

Learning path: Ontology → Epistemology → Ethics → Methodology → Hermeneutics → Synthesis → Praxis.

Each agent requires Methodology + Hermeneutics. Directors operate at Synthesis. Traders apply Praxis.

Prerequisites

Required Knowledge

Before learning Ghost’s system, you should understand:

Recommended Background

Helpful but not required:

• Experience reading charts (any timeframe)
• Familiarity with Fibonacci numbers
• Understanding of moving average types (SMA vs EMA)
• Exposure to volume analysis concepts

Learning Path

Phase 1: Foundation (Ontology)

Goal: Understand the conceptual vocabulary.

Study:

1. Read ONTOLOGY.md completely
2. Focus on Memory Types, Levels, and Confluence
3. Be able to explain each concept without reference

Checkpoint Questions:

• What are the three memory types and which agent reads each?
• What makes a level “strong”?
• What is the difference between horizontal and temporal confluence?
• What is an epoch and why does it matter?

Time: 1-2 hours of study

Phase 2: Single-Agent Mastery

Learn each agent in this order:

2A: Fib Agent (Structural Memory)

Why first: Fib provides the structural framework. Other agents operate within this framework.

Study sequence:

1. FIB_METHODOLOGY.md — Learn epoch identification
2. FIB_HERMENEUTICS.md — Learn the five questions
3. ghost/prompts/fib/lead.jinja2 — See how the agent thinks

Practice exercises:

1. Epoch identification: Open a 1-year chart. Identify 3 meaningful moves (epochs). Justify why each qualifies.
2. Level calculation: For one epoch, calculate all Fib levels (0%, 23.6%, 38.2%, 50%, 61.8%, 78.6%, 100%).
3. Multi-timeframe layering: Add weekly and daily epochs. Find where levels from different timeframes cluster.
4. Five questions drill: For the current price, answer all five hermeneutic questions in writing.

Checkpoint:

• Can you identify an epoch without second-guessing?
• Can you find confluence zones across timeframes?
• Can you explain why 61.8% is special?

Time: 3-4 hours

2B: MA Agent (Trend Memory)

Why second: MA shows how price moves within the Fib structure. It adds directionality.

Study sequence:

1. MA_METHODOLOGY.md — Learn stacking and clusters
2. MA_HERMENEUTICS.md — Learn the five questions
3. ghost/prompts/ma/analysis.jinja2 — See the agent output

Practice exercises:

1. Stack identification: On a daily chart, determine if the stack is bullish, bearish, or tangled.
2. Dynamic S/R: Identify which MA is currently acting as support or resistance.
3. Extension assessment: Calculate distance from 20 SMA and 50 SMA. Is price extended?
4. Monthly context: Add monthly MAs. What’s the long-term regime?

Checkpoint:

• Can you instantly recognize a bullish vs bearish stack?
• Can you identify MA clusters?
• Do you know when “extended” becomes “extreme”?

Time: 2-3 hours

2C: VWAP Agent (Positioning Memory)

Why third: VWAP reveals WHO is positioned where. This explains WHY levels hold or break.

Study sequence:

1. VWAP_METHODOLOGY.md — Learn session and anchored VWAPs
2. VWAP_HERMENEUTICS.md — Learn the five questions
3. ghost/prompts/vwap/analysis.jinja2 — See positioning analysis

Practice exercises:

1. Session VWAP analysis: For the last 5 sessions, note where price closed vs VWAP. Who controlled each session?
2. Trapped cohort identification: Find an anchored VWAP above price. Who bought there? How underwater are they?
3. Defending cohort identification: Find an anchored VWAP below price. Who bought there? How profitable are they?
4. Accumulation/distribution: Is the anchored VWAP slope rising or falling? What does this mean?

Checkpoint:

• Can you identify who is trapped and who will defend?
• Can you read the session regime from VWAP closes?
• Do you understand why VWAP reveals positioning, not just price?

Time: 2-3 hours

2D: Momentum Agent (Timing Layer)

Why last: Momentum tells you WHEN to act. It’s useless without structural context from other agents.

Note: Momentum is a timing layer, not a memory type. Unlike Fib/MA/VWAP which have both methodology and hermeneutics docs, momentum has only MOMENTUM_HERMENEUTICS.md. This is because:

• Methodology for momentum indicators (RSI, MACD, Keltner) is standard/universal — any trading reference covers calculation
• Hermeneutics (interpretation) is Ghost-specific — how to read these indicators for long-only trading

Study sequence:

1. MOMENTUM_HERMENEUTICS.md — The five questions for momentum interpretation
2. ghost/prompts/momentum/analysis.jinja2 — Full agent prompt with detailed rules
3. Practice RSI divergence identification
4. Practice MACD histogram reading
5. Practice squeeze detection

Practice exercises:

1. Divergence spotting: Find 3 examples of bullish divergence (price lower, RSI higher). What happened next?
2. Extreme percentile: When RSI hits bottom 5th percentile, what’s the typical forward return?
3. Squeeze detection: Identify a multi-timeframe squeeze. What happened when it released?
4. Confluence count: At current price, count bullish vs bearish momentum signals.

Checkpoint:

• Can you spot divergence quickly?
• Do you know when RSI is “extreme” vs just “oversold”?
• Can you read a squeeze setup?

Time: 2 hours

Quick Reference: The Five Questions (All Agents)

Each agent answers five hermeneutic questions. This table provides a quick reference without needing to open each *_HERMENEUTICS.md file.

Fib (Structural Memory)

MA (Trend Memory)

VWAP (Positioning Memory)

Momentum (Timing Layer)

Phase 3: Synthesis (Technical Director)

Goal: Learn to combine all four agents into unified analysis.

Study:

1. ghost/prompts/director/technical.jinja2 — The synthesis prompt
2. Focus on the “Synthesizing All Four” table
3. Understand confluence override rules

Practice exercises:

1. Full analysis: Run ghost technical VST and trace each conclusion to agent sources.
2. Conflict resolution: When agents disagree, which should you weight higher and why?
3. Citation practice: For each buy zone, write the [Agent] @ $price citations.
4. Conviction calibration: Given agent outputs, what conviction level is appropriate?

Checkpoint:

• Can you synthesize four agent reports into a coherent thesis?
• Do you know when confluence overrides historical accuracy concerns?
• Can you cite agent evidence properly?

Time: 2-3 hours

Phase 3B: Other Stacks (Reference)

The learning path above focuses on the Technical Stack. Ghost has two other synthesis layers:

These feed into the Market Strategist (final arbiter) alongside the Technical Director.

When to learn these:

• After mastering Phase 3 (Technical Director synthesis)
• News Stack: docs/NEWS_METHODOLOGY.md + docs/NEWS_HERMENEUTICS.md
• Market Structure: docs/MARKET_STRUCTURE_METHODOLOGY.md + docs/MARKET_STRUCTURE_HERMENEUTICS.md
• Options Analysis: docs/OPTIONS_HERMENEUTICS.md (critical for day trading)
• Market Strategist: Read ghost/prompts/director/strategist.jinja2

This learning path prioritizes the Technical Stack because it’s the foundation. News and positioning modify conviction but don’t replace structural analysis.

Phase 3C: Options Analysis (Day Trading)

Goal: Understand how dealer hedging creates intraday price levels and how the weekly cycle affects interpretation.

Study sequence:

1. ONTOLOGY.md — Options Structure Concepts section
2. OPTIONS_HERMENEUTICS.md — The Five Questions (Options)
3. MARKET_STRUCTURE_METHODOLOGY.md — How options data is gathered
4. MARKET_STRUCTURE_HERMENEUTICS.md — Full positioning interpretation

Practice exercises:

1. Day-of-week calibration: On Monday, note how max pain and gamma flip differ from Friday’s close. Why did they shift?
2. Dealer mode identification: For today, identify the dealer mode (BUILDING/ADJUSTING/PREPARING/UNWINDING) and adjust your interpretation weights accordingly.
3. Dual flip reading: Compare structural (ex-0DTE) vs tactical (0DTE) gamma flip. Where is the intraday battleground vs multi-day bias?
4. Five questions drill (Options): Answer all five hermeneutic questions from OPTIONS_HERMENEUTICS.md for the current ticker.
5. Friday session phases: On expiration day, note dealer behavior at 9:30, 11:30, 14:00, and 15:30. What changes in each phase?

Checkpoint:

• Can you explain why max pain has strong pull on Friday but weak pull on Monday?
• Do you know the difference between structural and tactical gamma flip?
• Can you identify which day-of-week rules apply today?
• Do you understand what happens during Friday’s hedge unwind (3:30pm)?

Time: 2-3 hours

Phase 4: Live Application

Goal: Apply the system to real trading decisions.

Practice:

1. Pre-market prep: Run full analysis before market open. Write your plan.
2. Intraday monitoring: Watch how price interacts with identified levels.
3. Post-session review: Compare what happened vs what you expected. Why did it differ?
4. Prediction tracking: Log predictions and verify them. Calculate your accuracy.

Feedback loop:

• After 10 predictions, review accuracy by agent source
• Identify which agents you over-trust or under-trust
• Adjust your synthesis weighting
• Note: Feedback is scoped to the current prompt regime — after a major prompt change, accuracy resets and needs 30-50 verified predictions to rebuild

Time: Ongoing

Common Mistakes

Beginner Mistakes

Intermediate Mistakes

Advanced Mistakes

Skill Progression Levels

Level 1: Novice

• Can identify basic levels (support/resistance)
• Knows what each indicator measures
• Can read agent outputs

Level 2: Practitioner

• Applies epoch criteria consistently
• Finds confluence across 2 agents
• Answers hermeneutic questions
• Makes predictions with reasoning

Level 3: Competent

• Synthesizes all 4 agents fluently
• Knows when to weight which agent
• Calibrates conviction appropriately
• Tracks and learns from predictions

Level 4: Proficient

• Reads positioning intuitively
• Anticipates level interactions
• Adjusts for market regime
• High accuracy on confluence zones

Level 5: Expert

• Develops new patterns
• Teaches others effectively
• Contributes to system improvement
• Consistently profitable application

Study Resources

Primary Documents (This Repository)

Practice Commands

# Full analysis (all agents)
ghost run VST

# Scout: scan watchlist for setups
ghost scout scan

# Scout: see what's on the watchlist
ghost scout list

# Technical stack only
ghost technical VST

# Verify predictions
ghost verify VST

# View prediction history
ghost predictions VST

Teaching Others

If you’re teaching this system to others:

Recommended Teaching Order

1. Start with ontology. Establish shared vocabulary.
2. Demo one agent. Show Fib analysis end-to-end.
3. Practice together. Identify epochs on the same chart.
4. Add agents incrementally. MA, then VWAP, then Momentum.
5. Synthesis last. Only after single-agent competence.

Effective Exercises

• Blind chart analysis: Student analyzes, then compare to Ghost output
• Five questions drill: Rapid-fire hermeneutic questions
• Citation practice: Student must cite agent evidence for every claim
• Prediction journal: Written predictions with reasoning, weekly review

What to Emphasize

• Concepts over tools
• Questions over answers
• Process over predictions
• Learning from errors

Related Documentation

Foundational Docs

• ONTOLOGY.md — What the concepts are
• EPISTEMOLOGY.md — How Ghost forms and validates knowledge
• ETHICS.md — What Ghost must and must not do

Per-Agent Docs

• *_METHODOLOGY.md — How to identify (per agent)
• *_HERMENEUTICS.md — How to interpret (per agent)
• OPTIONS_HERMENEUTICS.md — Day-of-week options interpretation

Practice Docs

• PRAXIS.md — Where analysis meets action
• INTERROGATION.md — Prompt signal density evaluation
• DAY_DESK_REFERENCE.md — Day desk tool catalog
• CLAUDE.md — System overview and commands