OpenClaw Stock Analysis

Core Problem & Solution

In the fast-paced A-share market, investors need to perform rapid, quantitative, and data-driven analysis on individual stocks to aid intra-day decision-making. Relying solely on subjective judgment or a single indicator can overlook bullish/bearish contradictions and struggle to systematically process market news sentiment.

The OpenClaw Stock Analysis system is built to address this. It establishes an automated analysis workflow from data acquisition to a composite score, combining complex multi-indicator technical analysis with quantified news sentiment to provide users with a structured analytical report. The core lies in its Composite Scoring Model (Technical 85% + News Sentiment 15%) and Arbitrage Space Recommendation Algorithm, designed to output quantified strength assessments and reference price levels for trading, rather than vague predictions.

Analysis Workflow & Key Steps

The system's workflow adheres to strict analytical discipline, broken down into key steps:

1. Data Acquisition & Segregation:

   • Historical Data: Uses akshare to fetch K-line data up to the last trading day's close, for calculating all technical indicators.
   • Real-time Data: Uses the Tencent API to obtain intraday real-time quotes (price, volume, etc.).
   • Key Point: The system explicitly differentiates between these two data types. All analysis is based on this distinction, and data timeliness is mandatorily declared in the output.

2. Technical Quantitative Assessment:

   • Calculates over 10 technical indicators from historical K-lines, including trend-based (MA, MACD), oscillator-based (KDJ, RSI, WR, CCI), volume-based (VR, CR, OBV), and channel-based (BOLL, BIAS) types.
   • Each indicator has defined scoring rules (e.g., KDJ < 20 is in oversold territory and adds points), culminating in a Technical Score with a full mark of 85 points.

3. News Sentiment Scoring:

   • Scrapes stock-specific news from East Money and performs sentiment scoring via keyword-level weighted matching (e.g., Strong Positive x3.0, Normal Positive x1.0, Weak Positive x0.5).
   • Incorporates temporal decay (100% for today, 50% for yesterday, 20% for the day before) and maps the news sentiment to a News Sentiment Score with a full mark of 15 points.

4. Composite Scoring & Output:

   • The Technical and News scores are added with their respective weights (total 100) to determine a Scoring Grade (e.g., ≥70 for Strongly Bullish).
   • Output strictly follows a "three-part structure": ① Data Facts (script output) → ② Technical Judgment (rule-based interpretation) → ③ Action Reference (conditional advice).
   • Must provide specific price levels such as support, resistance, and stop-loss (dynamically calculated based on BOLL lower band and 20-day low).

5. Batch Arbitrage Screening (V4.0+):

   • For a user-provided stock list, runs parallel calculations and ranks them using a dedicated Arbitrage Score Formula (Technical*40 + Upside Space*25 + News*10 + R/R*15 - Price Increase*10 + ...).
   • Outputs the TOP 10 recommendations, which include deep analysis such as "accessibility rating."

Applicable Scope & Important Caveats

This skill is not an "all-in-one" analytical tool; its design and operation have clear boundaries:

• Purely Technical & Sentiment-Driven: Analysis is based entirely on historical price data and news keyword matching. It does not include any fundamental analysis (e.g., PE, PB, ROE, revenue) and cannot obtain real-time capital flow data. Users should not inquire about such information.
• Analysis ≠ Decision: The output is a quantitative reference based on historical data and statistical rules, and it must never replace human judgment. Red lines include: strictly prohibiting predictions of specific price points, prohibiting the use of directive verbs (e.g., "buy"/"sell"), and requiring all conclusions to be framed as conditional advice (e.g., "suggests," "for reference").
• Model & Data Limitations: Indicators exhibit correlations (e.g., RSI and KDJ); news keyword matching is prone to misinterpretation (this must be noted in the output); technical indicators may become unreliable for limit-up/limit-down or suspended stocks. While dynamic stop-losses are superior to fixed percentages, they may still underperform in highly volatile sectors like the ChiNext.
• Mandatory Output Discipline: Every analysis must declare the data cutoff time, the analysis type, the limitations of the news sentiment score, and include a risk disclaimer (⚠️ The above analysis is for reference only and does not constitute investment advice.).