Stock trading bot: momentum rotation with a crash guard

An ML momentum-rotation strategy trading the Nasdaq-100, live on Alpaca paper trading since April 2026.

Strategy flow

1. Alpaca WebSocket: Nasdaq-100 quotes and daily bars, auto-reconnect with backoff.
2. Redis cache + Postgres: hot tick fan-out, 126 days of price history.
3. CatBoost momentum scorer: ranks the universe on 126-day lookback features.
4. Monthly rebalance planner: target weights, turnover-aware rotation.
5. Risk gates → Alpaca paper: crash guard, sector cap, idempotent orders, full audit trail.

Risk rules

Rule	What it does
Crash guard	Regime filter. When the index breaks trend, the book rotates to cash instead of chasing momentum down.
Sector cap	Concentration ceiling so a hot sector can’t own the whole portfolio.
Paper first	Live on Alpaca paper since April 2026. The real execution path, zero live capital until the track record earns it.
Monthly cadence	Rebalance is scheduled, not reactive. No intraday overtrading.

Tech stack

CatBoost, FastAPI, Redis, PostgreSQL, Alpaca WebSocket.

---

From the original write-up (April 2026)

Status at the time: Phase 10, 14-day paper trading validation. Core mission: Execute high-probability trades with “Defense in Depth” safety protocols.

The bot operated as a stateful graph where each node is a specialized agent, so market data, sentiment, and risk are analyzed in a strict, logical sequence.

The LangGraph pipeline

1. Scanner: ranks top tickers (AAPL, NVDA, etc.) via Claude based on momentum and volatility.
2. Market data: fetches OHLCV data and computes technical indicators (RSI, MACD, Bollinger Bands).
3. Research: aggregates news sentiment and fundamental analysis.
4. Strategy: Claude synthesizes data into a concrete signal (entry, SL, TP).
5. Risk gate: a 6-layer validation check (circuit breaker, position limits, R:R ratio).
6. Human gate: terminal-based approval interrupt for final execution authority.
7. Executor: automated order submission to the Alpaca API.
8. Journal: post-trade analysis saved to PostgreSQL for continuous learning.

“Defense in depth” safety system

Capital preservation is more important than profit.

Layer	Component	Function
Layer 1	Redis circuit breaker	Hard-stop on all trading if daily loss hits $150.
Layer 2	Paper guard	Hard-coded block that prevents live execution unless two separate flags are met.
Layer 3	Risk gate	Validates risk/reward is ≥ 2.0 and position size is capped at $500.
Layer 4	Human-in-the-loop	Requires manual approval in the terminal before any order hits the broker.

Backtesting

A dedicated backtesting engine (scripts/backtest.py) replicates live signal logic without LLM costs: long-only momentum, 3% stop loss / 6% take profit, with Sharpe ratio, profit factor and max drawdown tracked by the journal agent.

Graduation criteria (paper → live)

1. Zero unhandled exceptions over 14 days.
2. At least 20 full cycles completed.
3. Safety block rate remains below 30%.
4. 100% adherence to the $500 position limit.