Fengniu Limit-Up Capture Model

Problem

A limit-up close alone is not enough to infer next-day board continuation. This skill narrows the task to ranking today's limit-up pool after removing noisy names, using backtest-oriented signals around capital, board history, chart structure, and volume behavior.

How It Works

Fengniu Limit-Up scores limit-up stocks on a 100-point system:
- Capital strength (40 points): evaluates main net inflow, capital-flow ratio, and break above the EXPMA cost line, marked as the strongest predictor in the notes.
- Board-streak recognition (30 points): differentiates multi-board stocks, second boards, and pure first boards using board level and 30-day limit-up history.
- Technical pattern (20 points): checks bullish moving-average alignment, Bollinger upper-band breakout, and 60-day line breakout.
- Volume resonance (10 points): looks at volume ratio, turnover rate, and seal time.

Outputs are labeled as >=80 strong, 60-79 medium, and <60 weak, with automatic exclusion of ST names, stocks listed under 50 days, and names with two consecutive limit-downs within 10 days. The workflow uses akshare for limit-up pool, K-line, capital-flow, and company data, then computes with pandas and numpy to produce a terminal table, HTML report, and JSON file.

Boundaries

This is a short-term continuation-ranking tool, not a long-term valuation model. It includes discipline around sealed-board buying, pullback entry, the 2-2-2-4 position split, stop-loss, and take-profit rules, but live use still depends on order book, sector mood, and liquidity. The cited backtest covers 100 limit-up stocks from 2026-04-29 to 2026-04-30, with TOP20 average next-day gain of +4.94%; that result should not be generalized to all market regimes.