Upload ALPHA_FORGE_GUIDE.md

ALPHA_FORGE_GUIDE.md

@@ -0,0 +1,702 @@
+# AlphaForge: The Complete Guide (For Humans, Not Just Quants)
+
+## 🎯 What Is AlphaForge?
+
+**In one sentence:** AlphaForge is a robot that reads financial news, studies stock charts, predicts which stocks will go up, figures out how risky they are, then tells you exactly how much money to put in each one — all while tracking whether it's actually working.
+
+Think of it like having a team of 5 experts working for you 24/7:
+1. **The Chart Reader** (Alpha Model) — looks at price patterns
+2. **The News Junkie** (Sentiment Model) — reads all the headlines  
+3. **The Risk Calculator** (Volatility Engine) — measures danger
+4. **The Money Manager** (Portfolio Optimizer) — decides allocations
+5. **The Options Expert** (Options Pricer) — finds mispriced bets
+
+---
+
+## 📊 EVERY METRIC EXPLAINED LIKE YOU'RE 5
+
+### 1. Sharpe Ratio — "Am I Getting Paid for the Scary Parts?"
+
+**What it means:** For every $1 of risk you take, how many dollars do you get back?
+
+**Real-world analogy:**
+- You have two jobs:
+  - Job A: $50/hour, but you might get fired any day (scary!)
+  - Job B: $40/hour, super stable, health insurance
+- Job B has a better "Sharpe Ratio" — you get almost as much money with way less stress
+
+**The numbers:**
+| Sharpe | What It Means | Your Action |
+|--------|---------------|-------------|
+| Below 0.5 | You're barely getting paid for the stress | Stop or reduce by 50% |
+| 0.5 - 1.0 | Okay, but not amazing | Keep going, look for improvements |
+| 1.0 - 1.5 | Good! Professional territory | This is your baseline target |
+| 1.5 - 2.0 | Excellent | You're beating most hedge funds |
+| 2.0+ | Legendary | Renaissance Technologies level |
+
+**How it influences your trading:**
+- Sharpe drops below 0.5 for a month? → Reduce all position sizes by half
+- Sharpe stays above 1.5 for 3 months? → Increase sizes by 25%
+- Sharpe negative? → STOP. Something is broken.
+
+**Example:**
+```
+You make 15% return but your portfolio jumps up and down 20% (volatile)
+Sharpe = (15% - 4%) / 20% = 0.55  →  Meh
+
+Your friend makes 12% return but barely moves (calm)
+Sharpe = (12% - 4%) / 8% = 1.00  →  Better!
+```
+Your friend made LESS money but took WAY less risk. Sharpe says your friend won.
+
+---
+
+### 2. Sortino Ratio — "Am I Getting Paid for ONLY the Bad Scary Parts?"
+
+**What it means:** Like Sharpe, but it IGNORES good volatility (big upward jumps).
+
+**Real-world analogy:**
+- Imagine your monthly pay:
+  - Some months you get BONUSES (upward jumps)
+  - Some months you get FIRED (downward drops)
+- Sharpe punishes BOTH bonuses and firings
+- Sortino only punishes firings
+
+**The numbers:**
+| Sortino | What It Means |
+|---------|---------------|
+| Below 1.0 | Your losses are too big/frequent |
+| 1.0 - 1.5 | You're managing losses okay |
+| 1.5 - 2.5 | Strong loss control |
+| 2.5+ | Elite — you're a loss-cutting machine |
+
+**How it influences your trading:**
+- Sortino > Sharpe by a lot? → GREAT! Your upward jumps are huge. Keep doing what you're doing.
+- Sortino ≈ Sharpe? → Your wins and losses are similarly sized. Consider taking profits faster.
+- Sortino drops below 0.8? → Your stops are too loose. Tighten them immediately.
+
+**Example:**
+```
+Month 1: +2%  |  Month 2: +8%  |  Month 3: -3%  |  Month 4: +1%
+Sharpe sees: volatility = 4.5% → calculates based on ALL movement
+Sortino sees: only the -3% matters → calculates based on DOWNSIDE only
+
+Result: Sortino = 2.1, Sharpe = 0.9
+Conclusion: Most of your volatility is UPWARD. This is GOOD volatility.
+```
+
+---
+
+### 3. Information Coefficient (IC) — "Are My Predictions Actually Right?"
+
+**What it means:** If your model says "buy Apple, sell Tesla," did Apple actually go up and Tesla go down?
+
+**Real-world analogy:**
+- You have a weather app
+- Every day it says "rain" or "sun"
+- IC measures: when it said "rain," did it ACTUALLY rain?
+
+**The numbers:**
+| IC | What It Means | Can You Trade It? |
+|----|---------------|-------------------|
+| 0.00 | Random guessing | No |
+| 0.02 | Slightly better than coin flip | Yes, at huge scale ($1B+) |
+| 0.05 | Decent — most quant funds live here | Yes, this is solid |
+| 0.10 | Strong — you're genuinely predictive | Yes, excellent |
+| 0.15+ | Exceptional — top-tier signal | Yes, this is gold |
+
+**How it influences your trading:**
+- IC drops below 0.02? → Your signal died. STOP using it. Retrain the model.
+- IC is 0.05-0.08? → This is your bread and butter. Trade normal sizes.
+- IC is 0.10+? → Scale up. This signal is hot.
+- IC goes NEGATIVE? → Your model is now WRONG. Either STOP or FLIP the signal (do the opposite).
+
+**Example:**
+```
+Monday: Model says AAPL +2%, MSFT +1%, TSLA -1%
+Actual:   AAPL +3%, MSFT 0%,  TSLA -2%
+
+Ranking predicted: AAPL > MSFT > TSLA  ✓
+Ranking actual:    AAPL > MSFT > TSLA  ✓
+IC = 1.0 (perfect!)
+
+Tuesday: Model says TSLA +2%, AAPL +1%
+Actual:   TSLA -5%, AAPL +3%
+
+Ranking predicted: TSLA > AAPL  ✗
+Ranking actual:    AAPL > TSLA  ✗
+IC = -1.0 (terrible!)
+```
+
+---
+
+### 4. Max Drawdown — "What's the Worst Day of My Life Going to Feel Like?"
+
+**What it means:** The biggest peak-to-trough drop in your portfolio. 
+
+**Real-world analogy:**
+- You have $100,000
+- It grows to $130,000 (peak)
+- Then crashes to $110,000 (trough)
+- Max Drawdown = ($130K - $110K) / $130K = 15.4%
+
+**The numbers:**
+| Max DD | How You'll Feel | Survival Chance |
+|--------|-----------------|-----------------|
+| Under 5% | Barely notice | 99% |
+| 5-10% | Mildly annoyed | 95% |
+| 10-20% | Stressed but okay | 85% |
+| 20-30% | Panic setting in | 60% |
+| 30-50% | Sleepless nights | 30% |
+| Over 50% | You'll probably quit forever | 10% |
+
+**How it influences your trading:**
+- Set YOUR personal max drawdown BEFORE you start (AlphaForge default: 15%)
+- If you hit 50% of your max DD (7.5% if limit is 15%) → Reduce positions by 25%
+- If you hit 75% of your max DD → Reduce by 50%
+- If you hit 100% of your max DD → STOP. Reassess everything.
+
+**The Math That Hurts:**
+```
+Lose 10% → Need +11% to recover
+Lose 20% → Need +25% to recover  
+Lose 30% → Need +43% to recover
+Lose 50% → Need +100% to recover (DOUBLE your money!)
+Lose 90% → Need +900% to recover
+```
+This is why drawdown KILLS. Small losses compound into unrecoverable holes.
+
+---
+
+### 5. Calmar Ratio — "Am I Getting Paid Enough for My Worst Nightmare?"
+
+**What it means:** Your yearly return divided by your worst-ever drawdown.
+
+**Real-world analogy:**
+- You make $100K/year
+- But one year you lost $50K in a bad investment
+- Calmar = $100K / $50K = 2.0
+- Translation: "I make 2x my worst loss every year"
+
+**The numbers:**
+| Calmar | What It Means |
+|--------|---------------|
+| Below 1.0 | Your drawdown is BIGGER than your yearly return. You'll blow up eventually. |
+| 1.0 - 2.0 | Decent. You're getting paid for the risk. |
+| 2.0 - 3.0 | Strong. Your returns dwarf your worst losses. |
+| 3.0+ | Excellent. This is sustainable long-term wealth building. |
+
+**How it influences your trading:**
+- Calmar < 1.0? → EMERGENCY. Cut everything by 75%. This is a ticking time bomb.
+- Calmar 1.0-1.5? → Caution zone. Add diversification or reduce leverage.
+- Calmar 1.5-2.5? → Solid. This is where you want to live.
+- Calmar > 3.0? → You've found something special. Don't tell anyone.
+
+**Example:**
+```
+Trader A: 20% return, 15% max drawdown → Calmar = 1.33
+Trader B: 15% return, 5% max drawdown → Calmar = 3.0
+
+Trader A makes MORE but Trader B is the real winner.
+Trader B will compound wealth steadily without ever panicking.
+Trader A will have a 15% nightmare year and probably quit.
+```
+
+---
+
+### 6. Win Rate — "How Often Am I Right?"
+
+**What it means:** Percentage of trades/days that made money.
+
+**THE SECRET:** Win rate ALONE is meaningless.
+
+**Real-world analogy:**
+- Casino game 1: You win 70% of the time but when you lose, you lose 3x what you win
+- Casino game 2: You win 40% of the time but when you win, you win 3x what you lose
+- Game 2 is profitable. Game 1 will bankrupt you.
+
+**The numbers:**
+| Win Rate | With What Avg Win/Loss | Result |
+|----------|------------------------|--------|
+| 70% | Win $100, Lose $300 | LOSING (expectancy: -$20/trade) |
+| 50% | Win $120, Lose $100 | WINNING (expectancy: +$10/trade) |
+| 40% | Win $300, Lose $100 | WINNING (expectancy: +$20/trade) |
+| 30% | Win $400, Lose $100 | WINNING (expectancy: +$20/trade) |
+
+**How it influences your trading:**
+- Win rate drops 10% from your historical average? → Signal decay. Retrain.
+- Win rate > 60% but barely profitable? → Dangerous negative skew. One big loss wipes you out.
+- Win rate 45-55% with good profits? → The sweet spot. Balanced, sustainable.
+- Win rate < 40% but profitable? → You have positive skew. Mentally prepare for long losing streaks.
+
+**Example:**
+```
+You trade 100 times:
+- 40 wins, average win = $500  → Total wins = $20,000
+- 60 losses, average loss = $250 → Total losses = $15,000
+- Net profit = $5,000
+- Win rate = 40% but you're PROFITABLE!
+
+Key insight: It's not about being right. It's about making MORE when you're right
+than you lose when you're wrong.
+```
+
+---
+
+### 7. Profit Factor — "For Every $1 I Lose, How Much Do I Make?"
+
+**What it means:** Total profits divided by total losses.
+
+**Real-world analogy:**
+- You run a lemonade stand
+- Some days you make $200 (profits)
+- Some days you lose $100 (rainy days)
+- Profit Factor = $200 / $100 = 2.0
+
+**The numbers:**
+| Profit Factor | What It Means |
+|---------------|---------------|
+| Below 1.0 | You're losing money. Stop. |
+| 1.0 - 1.3 | Barely alive. One bad streak kills you. |
+| 1.3 - 1.8 | Healthy. Sustainable. |
+| 1.8 - 2.5 | Strong. Professional grade. |
+| 2.5+ | Excellent. Robust edge. |
+
+**How it influences your trading:**
+- PF < 1.0? → No edge. Stop immediately.
+- PF 1.0-1.3? → Fragile. Reduce size 50% until PF improves.
+- PF 1.3-1.8? → Solid. Keep doing what you're doing.
+- PF > 2.0? → Robust. You can survive multiple bad streaks.
+
+**Example:**
+```
+100 trades:
+- Winning trades totaled $80,000
+- Losing trades totaled $40,000
+- Profit Factor = 2.0
+
+For every $1 you lost, you made $2. Solid!
+
+But if transaction costs were $10,000:
+- Net wins = $70,000, Net losses = $50,000
+- Net PF = 1.4
+
+Transaction costs dropped your PF from 2.0 to 1.4!
+This is why keeping costs low is CRITICAL.
+```
+
+---
+
+### 8. Alpha — "Am I Actually Smart, or Just Riding the Market?"
+
+**What it means:** How much EXTRA return you made ABOVE what the market explains.
+
+**Real-world analogy:**
+- The whole stock market goes up 10% this year
+- If you also made 10%, your Alpha = 0%. You just rode the wave.
+- If you made 15%, your Alpha = 5%. You actually added value!
+
+**The numbers:**
+| Alpha | What It Means |
+|-------|---------------|
+| Negative | Underperforming a simple index fund. Just buy SPY. |
+| 0-2% | Matching the market. No edge. |
+| 2-5% | Good. You're adding real value. |
+| 5-10% | Strong. Significant skill. |
+| 10%+ | Elite. Top 1% of all managers. |
+
+**How it influences your trading:**
+- Alpha < 0% for 6+ months? → Just buy SPY. You're wasting time and money.
+- Alpha 0-3%? → You're a "closet indexer." Either commit to active or go passive.
+- Alpha 3-7%? → Solid edge. Keep refining but don't over-optimize.
+- Alpha > 7%? → Document everything. This edge won't last forever.
+
+**The Humbling Truth:**
+```
+80% of active managers have NEGATIVE alpha after fees.
+S&P 500 returns ~10%/year for decades.
+If you can't beat SPY after all this effort, just buy SPY and enjoy life.
+```
+
+---
+
+### 9. Beta — "How Much Do I Move When the Market Moves?"
+
+**What it means:** If SPY goes up 1%, how much does YOUR portfolio go up?
+
+**Real-world analogy:**
+- You're on a boat in the ocean
+- Beta = 1.0 → When waves rise 1 foot, your boat rises 1 foot (moves with the market)
+- Beta = 2.0 → When waves rise 1 foot, your boat rises 2 feet (amplified)
+- Beta = 0.0 → Your boat stays level regardless of waves (market-neutral)
+- Beta = -0.5 → When waves rise, your boat goes DOWN (inverse)
+
+**The numbers:**
+| Beta | When Market Goes Up 10% | When Market Goes Down 10% |
+|------|------------------------|--------------------------|
+| 0.0 | You don't care | You don't care |
+| 0.5 | You go up 5% | You go down 5% |
+| 1.0 | You go up 10% | You go down 10% |
+| 1.5 | You go up 15% | You go down 15% |
+| -0.5 | You go DOWN 5% | You go UP 5% |
+
+**How it influences your trading:**
+- Want MARKET EXPOSURE (bullish): Target Beta 0.8-1.2
+- Want LOW RISK (preserve wealth): Target Beta 0.3-0.6
+- Want MARKET NEUTRAL (hedge fund style): Target Beta 0.0 ± 0.1
+- Want INVERSE (betting on crash): Target Beta -0.5 to -1.0
+
+**Dynamic Beta Management (Advanced):**
+- Bull market detected? → Increase beta to 1.2-1.5 (capture upside)
+- Bear market detected? → Reduce beta to 0.3-0.5 (protect capital)
+- High volatility detected? → Reduce beta to 0.0-0.3 (survive first)
+
+**Example:**
+```
+March 2020: Market crashes 34%
+- Your Beta = 1.5 → You crash 51%. You're ruined.
+- Your Beta = 0.3 → You drop 10%. Annoying but survivable.
+- Your Beta = -0.2 → You GAIN 6.8%. You're the smart one.
+
+This is why Beta management is MORE important than stock picking.
+```
+
+---
+
+### 10. Information Ratio — "Am I Beating My Benchmark?"
+
+**What it means:** How much EXTRA return you get per unit of "different" risk from the benchmark.
+
+**Real-world analogy:**
+- You're a chef trying to beat Gordon Ramsay's restaurant
+- Gordon makes $1M/year with his standard menu
+- You make $1.2M/year with your creative menu
+- But your menu is risky — some dishes flop
+- Information Ratio = ($1.2M - $1M) / Risk of Being Different
+
+**The numbers:**
+| IR | What It Means |
+|----|---------------|
+| Below 0 | You're losing to the benchmark. Buy the index. |
+| 0-0.3 | Barely adding value. Not worth the effort. |
+| 0.3-0.5 | Decent. Some active management skill. |
+| 0.5-1.0 | Good. Meaningful skill. |
+| 1.0+ | Excellent. Top-tier active manager. |
+
+**How it influences your trading:**
+- IR < 0? → STOP active trading. Buy SPY and save yourself the stress.
+- IR 0-0.3? → You're "expensive index fund." Reduce active bets.
+- IR 0.3-0.7? → Solid. Keep your current strategy.
+- IR > 0.7? → You have genuine skill. Consider managing money for others.
+
+---
+
+### 11. Turnover — "How Much Am I Trading?"
+
+**What it means:** Percentage of your portfolio you trade per year.
+
+**Real-world analogy:**
+- You have a $100K portfolio
+- 50% turnover = You sold $50K worth and bought $50K worth of new stuff
+- 100% turnover = You sold everything and bought all new stuff
+
+**The numbers:**
+| Turnover | Style | Annual Cost at 3bps |
+|----------|-------|---------------------|
+| < 20%/year | Buy-and-hold | Almost nothing |
+| 20-50% | Long-term focused | Very low |
+| 50-100% | Moderate active | Low |
+| 100-200% | High active | Moderate |
+| 300%+ | Day trading | High |
+
+**How it influences your trading:**
+- Turnover > 150% but Sharpe < 1.0? → You're trading too much. Reduce frequency.
+- Turnover < 30% but Sharpe > 1.5? → Perfect. Low cost, high return.
+- Transaction costs > 20% of gross profits? → Your strategy is paying your broker, not you.
+
+**Example:**
+```
+You rebalance weekly. Each time you trade 40% of portfolio.
+Annual turnover = 40% × 52 weeks = 2,080%
+
+At 3bps cost per trade:
+Annual friction = 2,080% × 0.03% × 2 (buy + sell) = 1.25%
+
+If your gross alpha is 3%, your NET alpha is 3% - 1.25% = 1.75%
+You lost 42% of your edge to transaction costs!
+
+SOLUTION: Rebalance monthly instead.
+Annual turnover = 40% × 12 = 480%
+Annual friction = 480% × 0.03% × 2 = 0.29%
+Net alpha = 3% - 0.29% = 2.71%
+
+You DOUBLED your net alpha just by trading less!
+```
+
+---
+
+## 🧠 HOW EACH ALPHA FORGE COMPONENT WORKS
+
+### Component 1: The Alpha Model (The Chart Reader)
+
+**What it does:** Looks at 60 days of price/volume data and predicts if a stock will go up.
+
+**How it works (simple):**
+1. Every day it calculates: RSI, MACD, moving averages, volatility, returns
+2. It feeds these into 3 "brains":
+   - **LSTM Brain**: "I see a pattern! Every time RSI hits 70 then drops, the stock falls"
+   - **Transformer Brain**: "Day 15 and Day 42 were the most important days. They had unusual volume"
+   - **XGBoost Brain**: "If volume > average AND price breaks 20-day high, 73% chance of +2%"
+3. It averages all 3 opinions into one prediction
+
+**What you get:** "Apple predicted +2.3% over next 5 days"
+
+**How to use it:**
+- Positive prediction = Consider buying
+- Negative prediction = Consider selling/shorting
+- Check IC to see if predictions are actually working
+
+---
+
+### Component 2: The Sentiment Model (The News Junkie)
+
+**What it does:** Reads financial headlines and converts the "mood" into a number.
+
+**How it works (simple):**
+1. Takes a headline: "Apple beats earnings expectations by 20%!"
+2. FinBERT reads it and thinks: "This is VERY positive"
+3. Converts to score: +0.95 (on -1 to +1 scale)
+4. Does this for ALL news about ALL stocks
+5. Aggregates: "Apple had 5 positive, 2 negative articles today. Net sentiment: +0.62"
+
+**What you get:** "Apple sentiment today: +0.62 (bullish)"
+
+**How to use it:**
+- Sentiment > +0.5? → Bullish signal. Good time to buy.
+- Sentiment < -0.5? → Bearish signal. Consider selling.
+- Sentiment near 0? → No news edge. Rely on other signals.
+
+---
+
+### Component 3: The Volatility Engine (The Risk Calculator)
+
+**What it does:** Predicts how much a stock is going to jump around.
+
+**How it works (simple):**
+1. **GARCH Model** (classic): "If yesterday was wild, today will probably be wild too"
+2. **LSTM Model** (neural): "I see complex patterns in how volatility clusters"
+3. Combines both into: "Apple's volatility next week: 22% annualized"
+4. Also builds a **covariance matrix**: "When Apple crashes, Microsoft usually drops 60% as much"
+
+**What you get:** 
+- Individual stock volatility forecasts
+- Correlation matrix (how stocks move together)
+- Risk warnings: "Portfolio volatility is 18% but your target is 10%. Reduce positions!"
+
+**How to use it:**
+- High predicted volatility? → Reduce position size (you'll get stopped out)
+- Low predicted volatility? → Can increase position size
+- Correlations spiking? → Diversify more (everything is moving together)
+
+---
+
+### Component 4: The Portfolio Optimizer (The Money Manager)
+
+**What it does:** Takes predictions and risk, then decides EXACTLY how much to put in each stock.
+
+**How it works (simple):**
+1. Gets predictions: "Apple +2.3%, Microsoft +1.1%, Tesla -0.5%"
+2. Gets risk: "Apple vol 22%, Microsoft vol 18%, correlation 0.6"
+3. Solves: "How do I maximize return while keeping risk low and not putting too much in one stock?"
+4. Outputs: "Put 12% in Apple, 10% in Microsoft, 0% in Tesla, rest in cash/others"
+
+**It includes real-world constraints:**
+- Max 20% in any single stock (don't put all eggs in one basket)
+- Transaction costs (every trade costs money)
+- Turnover penalty (don't trade too much)
+
+**What you get:** Exact percentage allocations for every stock.
+
+**How to use it:**
+- Follow the allocations exactly
+- Rebalance when it says to (usually every 5 days)
+- Don't override it with your "gut feeling" — the math is smarter than you
+
+---
+
+### Component 5: The Options Pricer (The Derivatives Expert)
+
+**What it does:** Predicts fair prices for stock options and finds mispriced ones.
+
+**How it works (simple):**
+1. Takes option details: "Apple $180 call, expires in 30 days, Apple at $175"
+2. Black-Scholes formula says: "Fair price is $5.20"
+3. Neural network says: "Actually, fair price is $6.10 (market is messier than the formula thinks)"
+4. Market price is $4.80
+5. Alert: "UNDERPRICED by 27%! Buy signal!"
+
+**What you get:**
+- Fair option prices
+- Mispricing alerts (buy underpriced, avoid overpriced)
+- Implied volatility predictions
+
+**How to use it:**
+- Underpriced options? → Buy them or sell puts (collect premium)
+- Overpriced options? → Sell them or buy puts (hedge)
+- Use as hedging: "Buy cheap puts to protect your stock portfolio"
+
+---
+
+### Component 6: The Backtest Engine (The Time Machine)
+
+**What it does:** Pretends to trade using historical data to prove the strategy works.
+
+**How it works (simple):**
+1. Rewinds to January 2020
+2. Every day: gets predictions → runs optimizer → simulates buying/selling
+3. Includes REAL costs: 0.03% commission + slippage
+4. Tracks: money made, worst losing streak, how often right
+
+**What you get:**
+- "If you traded this way from 2020-2024, you'd have made 23%/year"
+- "Worst drop was 12%. Can you handle that?"
+- "You were right 52% of the time but winners were 2x bigger than losers"
+
+**How to use it:**
+- Sharpe < 1.0 in backtest? → Don't trade it with real money
+- Max drawdown > your pain threshold? → Reduce leverage or don't trade it
+- Backtest looks amazing but IC is 0.01? → It's curve-fitted. It will fail.
+
+---
+
+### Component 7: Regime Detection (The Market Mood Ring)
+
+**What it does:** Figures out if the market is in "bull mode," "bear mode," or "panic mode."
+
+**How it works (simple):**
+1. Looks at SPY (S&P 500) every day
+2. Calculates: Is price going up? Is volatility high?
+3. Labels:
+   - **Bull**: Price trending up, calm → Aggressive mode
+   - **Bear**: Price trending down → Defensive mode
+   - **High Vol**: Everything jumping around → Survival mode
+
+**What you get:** "Current regime: BULL. Strategy has returned +35%/year in this regime."
+
+**How to use it:**
+- Bull regime? → Increase beta, take more risk
+- Bear regime? → Reduce beta, increase cash, buy puts
+- High vol regime? → Reduce everything, wait it out
+
+---
+
+## 🔗 THE FULL PIPELINE (How It All Connects)
+
+```
+6:00 AM — Your day starts
+    |
+    ├── 📊 Market Data (yesterday's prices, volume)
+    │      └── Calculate: RSI, MACD, moving averages, returns
+    │      └── Alpha Model: "Apple +2.3%, MSFT +1.1%, TSLA -0.5%"
+    │
+    ├── 🗞️ Overnight News (headlines, earnings, macro)
+    │      └── FinBERT reads them
+    │      └── Sentiment Model: "Apple sentiment +0.62, TSLA -0.31"
+    │
+    │              ↓
+    │      Combined Alpha = 70% × Price Alpha + 30% × Sentiment Alpha
+    │      "Apple final score: +1.8 (strong buy)"
+    │
+    ├── 📉 Volatility Engine
+    │      └── "Apple risk = 18%, MSFT risk = 12%, correlation = 0.6"
+    │      └── Covariance matrix built
+    │
+    │              ↓
+    │      🧮 Portfolio Optimizer
+    │      "Given these predictions and risks, best allocation:"
+    │      Apple: 12% | MSFT: 10% | NVDA: 8% | Cash: 70%
+    │      (Conservative because volatility is elevated)
+    │
+    │              ↓
+    │      ⚡ Options Engine (optional)
+    │      "Apple $180 calls underpriced by 8% → buy 5 contracts"
+    │      "Use as hedge against tech exposure"
+    │
+    │              ↓
+    │      📈 Backtest Check
+    │      "This exact allocation pattern has 1.4 Sharpe historically"
+    │      "Max drawdown in similar conditions: 11%"
+    │      "Proceed with confidence"
+    │
+    9:30 AM — Market opens
+    │      └── Execute orders (limit orders, not market!)
+    │      └── Log everything
+    │
+    4:00 PM — Market closes
+    │      └── Record PnL
+    │      └── Update metrics
+    │      └── If down 6% this month → Reduce positions 50%
+    │
+    10:00 PM — System runs overnight
+    │      └── Re-train models with new data
+    │      └── Update volatility forecasts
+    │      └── Prepare tomorrow's signals
+```
+
+---
+
+## 🏆 YOUR PATH TO GOAT STATUS
+
+### Stage 1: Paper Trading (Months 1-3)
+- Run AlphaForge on paper money
+- Track ALL metrics daily
+- Don't trade real money until Sharpe > 1.0 consistently
+
+### Stage 2: Small Real Money (Months 4-6)
+- Trade 10% of intended capital
+- Focus on ONE alpha source until you master it
+- Journal every trade
+
+### Stage 3: Scale Up (Months 7-12)
+- Increase to 50% capital
+- Add second uncorrelated alpha source
+- Implement dynamic position sizing
+
+### Stage 4: Full Deployment (Year 2+)
+- Full capital allocation
+- Options hedging for tail risk
+- Tax-loss harvesting
+
+### Stage 5: GOAT (Year 3+)
+- Sharpe > 1.5 consistently
+- Max drawdown < 15%
+- Alpha > 5% annual
+- You're now in the top 5% of all active managers
+
+---
+
+## ⚠️ THE RULES THAT SEPARATE GOATS FROM BLOW-UPS
+
+1. **Survival First, Profit Second** — Compounding only works if you don't blow up
+2. **Process Over Outcome** — A good process with a bad result is variance, not failure
+3. **Never Risk > 2% on One Position** — Even the "sure thing" can gap down 50%
+4. **Never Add to Losers** — Averaging down is how $1M accounts become $0
+5. **Sleep Well or Reduce Size** — If you can't sleep, your risk is too high
+6. **Check PnL Once Per Day Max** — Intraday noise creates bad decisions
+7. **25% Drawdown = Mandatory Break** — Reassess everything. Don't revenge trade.
+8. **AlphaForge Is Smarter Than Your Gut** — The math removes emotion. Trust it.
+
+---
+
+## 📚 ESSENTIAL READING
+
+1. **"A Random Walk Down Wall Street"** — Burton Malkiel (understand markets)
+2. **"The Black Swan"** — Nassim Taleb (understand tail risk)
+3. **"Advances in Financial Machine Learning"** — Marcos Lopez de Prado (technical depth)
+4. **"Inside the Black Box"** — Rishi Narang (how quant funds work)
+
+---
+
+**Remember:** The goal isn't to get rich quick. The goal is to compound wealth steadily while you sleep well at night. AlphaForge gives you the tools. Discipline gives you the results.
+
+**Now go build your edge.** 🐐