Stock Dragon-Tiger List Analysis System

Problem

Dragon-Tiger list data often shows only who bought and who sold, but it is hard to infer which market sentiment stage the short-term market is in, which names fit the current cycle, and how aggressive position sizing should be. When limit-up counts, consecutive-board height, and institutional/hot-money seat resonance change quickly, ranking by activity alone can mix strong setups with late-cycle decay.

How It Works

The skill separates the workflow into two layers: identify the sentiment cycle first, then classify individual stocks.

• Data ingestion uses akshare to pull the limit-up pool, limit-down pool, Dragon-Tiger list summary, individual stock seat details, and the SSE Index, supporting metrics such as consecutive-board height, loss-effect pressure, seat structure, and broad-market bias.
• The analysis layer outputs a four-stage sentiment framework (ice, recovery, climax, retreat), a 0–100 sentiment thermometer, A+/A/B+/B/C/D classifications, and seat-resonance types such as institutional, hot-money, and northbound participation or institutional-only buying.
• Recommendations adapt by stage: during climax periods, some B+/B/C names may be downgraded to avoid; during ice periods, position sizing is capped. It generates an HTML report while preserving the original institutional list, hot-money list, and composite data tables.

Boundaries

It is better suited for post-close short-term review and next-day candidate screening than intraday real-time decisions. Dragon-Tiger data usually updates after 16:30 on trading days; classifications are based on historical backtests and do not guarantee future returns. Signals such as institutional-only buying may suffer from small-sample bias, so stop-losses, position limits, and liquidity constraints should still be applied.