Bar-by-Bar K-line Analysis (逐K量价分析)

Purpose

Provide detailed, candle-by-candle analysis of daily K-line data using the Wyckoff and Volume-Price (量价操盘) framework. The final output is a self-contained interactive HTML report with six mandatory sections: header stats, price + volume charts, K线全景总结 (phase panorama summary), 未来走势推演 (scenario probability analysis), key levels matrix, and per-candle detailed table.

When to Use

Activate when the user requests:

"逐K分析"、"逐根分析" — bar-by-bar analysis of every candle
"一根一根分析"、"每个K线都看看" — detailed per-candle breakdown
"详细分析每一根K线"、"把每根都过一遍" — exhaustive candle examination
Any request that implies looking at individual daily candles
"分析XX股票的K线"、"帮我看看XX最近走势"

Do NOT use for:

Brief overviews or summary trend analysis (use the broader Wyckoff skills directly)
Single-candle questions (eg. "这根K线什么意思")
Minute-level or weekly-level analysis (this skill is optimized for daily K-lines)

Prerequisites

Before analysis, load the Wyckoff and Volume-Price foundational skills. Invoke at minimum: 量价操盘1 (Three Laws), 量价操盘3 (Volume Spikes), 量价操盘4 (Shadows), 量价操盘5 (WRB/NR), 量价操盘6 (Entry Check / ND-NS), 量价操盘7 (Bullish Signals), 量价操盘8 (Bearish Signals), 威科夫4 (Accumulation), 威科夫5 (Distribution).

Data retrieval: use westockdata skill for A/HK/US stocks/indices, or mcp__tdx-connector tools. For global indices not covered by these sources (eg. KOSPI, Nikkei), use East Money API (push2his.eastmoney.com/api/qt/stock/kline/get). Minimum 50 candles needed for meaningful phase segmentation.

Analysis Workflow

Step 1: Data Collection

Retrieve daily K-line data (OHLCV) for the target with at least 50-60 candles. Parse into chronological order (oldest first).

Step 2: Candle Classification

For each candle, compute and label:

| Dimension | Check | Label | |-----------|-------|-------| | Body direction | close >= open? | 阳(bullish) / 阴(bearish) | | Range width | (h-l) vs 10-day avg | WRB: >1.5x avg / NR: <0.6x avg | | Upper shadow | h-max(o,c) > body0.5 | UT / 上影线 | | Lower shadow | min(o,c)-l > body0.5 | SPRING / 下影线 | | Body strength | body/(h-l) | 光头(>0.85) / 正常 / 十字(<0.15) |

Step 3: Volume Analysis

Compare volume against 10-day MA and prior day:

| Level | Threshold | Tag | |-------|-----------|-----| | 天量 | >2.0x 10d MA | 天量 | | 放量 | >1.3x 10d MA | 放量 | | 正常 | 0.7-1.3x 10d MA | 正常 | | 缩量 | 0.5-0.7x 10d MA | 缩量 | | 极缩 | <0.5x 10d MA | 极缩 |

Step 4: Wyckoff Signal Mapping

Map each candle to Wyckoff signals by combining range, direction, volume, and structural position. Reference references/signal_map.md for the complete mapping table.

Signal priority: SC/BC/SOW > SOS/JOC/UTAD > UT/SPRING/NS > AR/ST/LPSY > NR/休整

Step 5: Phase Segmentation

Divide the candle sequence into phases. See references/signal_map.md for triggers and colors.

Step 6: K线全景总结 (Panorama Summary)

This is a mandatory section. For each identified phase, produce a summary card with:

Phase number + name + date range (eg. "阶段四：JOC加速上涨 (5/4-5/11)")
The boundary trigger candle (which candle started/ended this phase and why)
2-4 sentence narrative connecting key candles within the phase
Quantitative mentions: exact price points, volume numbers, percentage changes
Use bold tags <b> for Wyckoff signal names (SOS/JOC/BC/SOW/NS/ND/UT/SPRING)
Mark dangerous phases with class="danger" on the summary-card div

Writing style: be specific with numbers, name every signal, connect the evolving narrative. Each phase card should be 60-120 words.

Step 7: 未来走势推演 (Scenario Probability Analysis)

This is a mandatory section. Generate four probability-weighted scenarios:

| Scenario | Probability Range | Type | |----------|:------------------:|------| | A: Bullish continuation | 25-40% | scn-bull | | B: Range-bound / distribution | 20-35% | scn-neutral | | C: Bearish / SOW confirmation | 20-30% | scn-bear | | D: Crash / black swan | 5-15% | scn-crash |

Each scenario card must contain:

Descriptive title with emoji marker
Probability percentage (must sum to 100% across all 4)
Path description: step-by-step narrative of how this scenario unfolds
3 supporting signals: numbered list of existing evidence that supports this scenario
Confirmation conditions: specific, falsifiable trigger conditions that would raise confidence

After the four scenarios, add:

Key level matrix: 8-cell grid showing resistance, support, current, and target levels
Bull/Bear evidence weight matrix: two-column layout listing evidence with % weights
Comprehensive judgment box: overall tilt, core observation variable, and operation advice

Probabilities must be grounded in the signals found in the data, not arbitrary. Reference the specific candles/signals from the analysis.

Step 8: Per-Candle Detail Table

Generate a scrollable HTML table with these columns: date, open, close, high, low, volume, 阴/阳, K线类型 (WRB/NR+UT+SPRING), 量级 (volume label), 信号标签, 威科夫/量价分析 (analysis text).

Step 9: Generate HTML Report

Produce a self-contained HTML file using the template in assets/report_template.html as structural reference. The report must include ALL of:

Header: Stock name, date range, framework, stat boxes (total range %, latest close, candle count, max, min, phase count)
Price chart (Chart.js line, red fill): closing price overlaid with phase-colored background
Volume chart (Chart.js bar): red/green colored by close vs prev close
Phase legend: colored tags for each identified phase
Phase panorama summary (Step 6 output): grid of phase summary cards
Scenario probability analysis (Step 7 output): four scenario cards + key levels + evidence matrix
Per-candle table (Step 8 output): scrollable detailed table
Signal tag legend: 多(bull)/空(bear)/警(warn)/观(observe)/阶段(phase) with colored tag examples

Mandatory visual conventions (Chinese market):

Price up (涨) → red (#c62828), Price down (跌) → green (#1b5e20)
Volume bars: red if close >= prev_close, green if close < prev_close
Signal tags CSS classes: .tag-bull (green bg), .tag-bear (red bg), .tag-warn (orange bg), .tag-obs (blue bg), .tag-phase (purple bg)
Highlight rows: .highlight-sc, .highlight-sos, .highlight-bc, .highlight-sow, .highlight-row (for JOC)

Table row highlighting (apply CSS class based on signal tags):

Contains 'SOW' → highlight-sow
Contains 'BC' → highlight-bc
Contains 'SOS' → highlight-sos
Contains 'JOC' → highlight-row

Analysis cell writing style:

Each cell: 50-100 characters, structured as: "[K线形态描述] → [威科夫信号判断] → [操作含义]"
Be specific: cite exact numbers (eg. "量14.2万手，较前日增+22%")
Use ★ for phase-defining candles (SC, SOS, JOC, BC, SOW)
Implication: what this candle means for the ongoing phase
Reference prior candles to build narrative continuity

CSS requirements — All the CSS classes defined in assets/report_template.html MUST be included. Do not skip any class. This includes: base styles, card styles, scenario grid styles, summary grid styles, table highlight styles, tag styles, signal item styles, key-value grid styles, and evidence matrix styles.

Output

A single HTML file named {stock_name}_逐K分析.html saved to the current workspace directory. Present via present_files tool after completion.

CRITICAL: Do NOT skip Sections 6 (K线全景总结) and 7 (未来走势推演). Every report must include both.

Notes

This is an exhaustive, detailed analysis. Each candle matters.
The analysis is objective — report what the volume and price action show.
When the data spans holidays, mark gap days in the table.
Use 量 (volume) short for 成交量 in the table to save space.
The scenario probabilities must be internally consistent and grounded in detected signals.
For international indices not in westockdata/TDX, use East Money global index API (secid=100.{code}, eg. KS11 for KOSPI, N225 for Nikkei, FTSE for FTSE100).