200-Day Trend Filter — AI Quant Strategy Skill

Description

Turn daily price data into a trend research report

An original Lingmirage AI skill with explicit long-or-cash rules and a standalone Python backtester. Download includes SKILL.md and scripts/quant_backtest.py.

Strategy rules

Signal: adjusted close above its trailing 200-session average. Execute at the following close. Long-only, no leverage; otherwise cash. Default trading friction: 10 bps per side, including final exit.

Inputs and outputs

Supply a licensed CSV with Date and AdjustedClose, at least 300 rows. Python 3.9+ required. Outputs JSON with total return, CAGR, max drawdown, zero-risk-free Sharpe, entries, equity curve and data fingerprint. No market data or AI subscription included.

Historical backtest — not live performance

SPY adjusted daily close, Yahoo Finance; 2017-06-21 to 2026-09-04 after a 200-session warm-up. Long/cash, no leverage, next-close execution, cash yield 0%, 10 bps each way including final liquidation. Total return +141.51%; CAGR 10.05%; maximum drawdown -24.97%. Same-period gross SPY buy-and-hold: approximately +264.75%, higher than this strategy.

Retrospective single-market sample, not independent out-of-sample validation. Historical results do not predict future returns. Market data and AI subscriptions are not included.

After purchase

Once payment is confirmed, open My Account → Downloads and download the ZIP. Unzip it, read SKILL.md, and use scripts/quant_backtest.py with Python 3.9+ and your licensed price CSV. This is a research skill, not automated trading or investment advice.

What is inside your ZIP?

• SKILL.md: readable strategy rules and the research workflow for you or an AI coding assistant.
• scripts/quant_backtest.py: standalone Python backtester. No extra Python packages or broker API keys required.

Is this right for you?

For self-directed learners who want to inspect rules and test their own data. You should be comfortable extracting a ZIP, preparing a CSV and running a Python command. An AI coding assistant is optional. This is not a hosted app, live signal feed or automatic trading bot.

Before you buy

Requires Python 3.9+ and at least 300 rows of licensed, split/dividend-adjusted daily prices. Data and AI subscriptions are not included. You can read the skill on mobile or tablet; running the backtest needs a suitable Python environment.

Read the setup guide · Find purchased downloads