Forge Trader

• Alpaca Market Data — Real-time quotes, bars (1m/5m/1d), trade stream via WebSocket
• News & Sentiment — Web search, RSS feeds, earnings calendars, SEC filings
• Technical Indicators — RSI, MACD, Bollinger Bands, VWAP, moving averages (computed locally)
• Alternative Data — Social sentiment, insider trading filings (future phase)

• LLM-powered market condition analysis (macro, sector, individual stock)
• Pattern scanning against active strategy criteria
• Conviction scoring: each opportunity rated 1-10 with supporting rationale
• Research outputs stored in SQLite for backtesting and review

• Modular strategy framework — each strategy is a Python class with scan(), signal(), size()
• Multiple strategies can run concurrently with portfolio-level allocation
• Backtesting harness using historical Alpaca data
• Strategy performance tracking (Sharpe, Sortino, max drawdown per strategy)

• Alpaca REST API for order submission (market, limit, stop-limit, trailing stop)
• WebSocket for real-time fill confirmations and position updates
• Pre-flight validation: risk checks must pass before any order
• Smart order routing: limit orders with timeout → market if not filled

See Risk Management tab for full details.

• SQLite DB — Every trade, fill, P&L event, daily portfolio snapshots
• This Portal — Live dashboard with performance metrics (future: real-time)
• Teams Alerts — Daily summary, trade notifications, risk alerts
• Monthly Reports — Performance attribution, strategy breakdown, cost analysis

Thesis

Large-cap stocks that deviate significantly from their short-term mean tend to revert. Buy deeply oversold, sell on confirmed reversion, with risk controls tuned by 12-month backtest.

Entry Criteria

• RSI(14) below 25 on daily chart (truly oversold only)
• Price below lower Bollinger Band (20, 2σ)
• No pending earnings within 5 trading days
• Average daily volume > 1M shares
• Market cap > $10B (avoid small-cap traps)

Exit Criteria

• Profit target: +8% gain (let winners run)
• RSI exit: RSI crosses above 55 (confirmed reversion)
• Stop loss: -3% from entry (wider stop to avoid chop)
• Time stop: Close after 7 trading days

Backtest Results (March 2025 – March 2026)

Sharpe: 1.58 · Win Rate: 61.1% · Profit Factor: 2.10 · Max DD: 1.8% · Return: +6.82%

90 trades over 12 months (~7.5/month). Avg win: +4.71%. Avg loss: -3.44%. Avg hold: 5.4 days.

Why These Parameters

Tested 15 parameter combinations. Key finding: the baseline 2% stop was too tight — getting stopped out on normal volatility before stocks could revert. RSI<25 (vs 30) filters to only deeply oversold setups with higher conviction. 8% profit target captures the full reversion move instead of cutting early.

Thesis

Stocks breaking out of consolidation patterns with volume confirmation tend to continue. Ride the trend with trailing stops.

• Price breaks above 20-day high with 1.5x average volume
• MACD histogram positive and increasing
• Trailing stop: 2x ATR(14) below current price
• Complements mean reversion — works in trending markets

Thesis

LLM analysis of earnings, FDA decisions, macro events, and sentiment shifts provides edge in predicting post-event price moves. This is where AI gives the biggest advantage.

• Pre-earnings positioning based on sentiment + estimate analysis
• Post-announcement gap analysis (buy the dip on overreactions)
• Macro catalyst plays (Fed decisions, inflation data)
• Real-time news processing for rapid-reaction trades

Thesis

24/7 crypto markets are ideal for an always-on AI agent. Higher volatility = more opportunities, but requires tighter risk controls.

• BTC/ETH swing trades on 4h/daily timeframes
• Altcoin momentum plays during crypto bull cycles
• Pairs trading (BTC vs ETH relative strength)
• DeFi yield opportunities (future consideration)

Covered Calls

• Sell OTM calls against existing stock positions for premium income
• Target 30-45 DTE, 0.20-0.30 delta (70-80% OTM)
• Roll or close at 50% profit or 21 DTE, whichever comes first
• Only on positions where I'm comfortable capping upside
• Expected yield: 1-3% monthly on covered positions

Cash-Secured Puts

• Sell puts on stocks I want to own at a lower price — get paid to wait
• Target 30-45 DTE, 0.20-0.30 delta (support levels)
• Strike = price I'd buy the stock at anyway based on research layer analysis
• If assigned: great, I wanted the stock. If not: keep the premium.
• Prerequisite: Only on stocks that pass the full research pipeline

Why Start Here

Both strategies are defined-risk by nature — covered calls risk is just capped upside, cash-secured puts risk is owning the stock (which you already wanted). No naked exposure, no margin blowup scenarios. This is how institutions generate yield.

Vertical Spreads (Bull Put / Bear Call)

• Use for event-driven plays where LLM analysis provides conviction
• Max loss = spread width minus premium received (known at entry)
• Earnings plays: sell put spread if bullish, sell call spread if bearish
• Typical structure: 5-wide spread, collect 30-40% of width as premium
• Key edge: LLM sentiment analysis determines direction, options define exact risk/reward

Long Debit Spreads

• Cheaper directional exposure for high-conviction catalyst plays
• Buy ATM call + sell OTM call = capped cost, leveraged upside
• Use when risk/reward on stock alone isn't compelling enough
• Max risk = debit paid (fully defined)

Iron Condors / Iron Butterflies

• Range-bound plays on low-volatility large caps
• Sell both put spread + call spread = profit if price stays in range
• Complements mean reversion thesis with premium collection

Calendar Spreads

• Exploit IV crush post-earnings: sell front-month high IV, buy back-month
• Profit from time decay differential

LEAPS (Long-Term Equity Anticipation Securities)

• 12-24 month deep ITM calls as stock replacement (0.80+ delta)
• Capital efficient way to hold long-term positions
• Frees up cash for other strategies

Gate

Stage 3 only unlocks after: 6+ months live trading, positive Sharpe ratio, and Tim's explicit approval. These strategies have more moving parts and require proven Greeks management.

• No naked options. Ever. Every position must be defined-risk.
• No selling strangles/straddles without protective wings (iron condor/butterfly only).
• Max options allocation: 30% of portfolio in options positions
• Max single options trade: 5% of portfolio at risk
• Always know max loss before entry. If max loss isn't defined, the trade doesn't happen.
• No earnings plays until Stage 2+ with proven directional accuracy
• Assignment management: Close spreads at 21 DTE to avoid early assignment risk
• Liquidity requirement: Only trade options with bid-ask spread < 10% of premium

• Trade notifications: Every trade → Teams message to Tim
• Daily summary: End-of-day P&L, positions, risk metrics
• Kill switch: Tim can halt all trading with one command
• Approval threshold: TBD — trades above $X may require Tim's approval
• Weekly review: Forge prepares performance report for Tim every Sunday