| name | guanlan-market-map |
| description | Analyze stocks, crypto, indices, or other liquid markets with the four-layer Guanlan framework: MA50/MA200 for regime and direction, MACD for momentum, TD Sequential for exhaustion, and Bollinger Bands for volatility state. Use when the user asks for technical analysis, chart diagnosis, bull/bear regime classification, entry/exit scenarios, indicator confluence, risk plans, or an explanation of MA, MACD, TD Sequential, and Bollinger signals. Accept charts, OHLCV data, indicator readings, or a ticker plus timeframe. Produce conditional scenarios and invalidation levels, not deterministic predictions or personalized financial advice. |
Guanlan Market Map
Turn four familiar indicators into one state-based decision system. Diagnose the environment before interpreting a signal; never let an isolated indicator dictate a trade.
Read references/framework.md before performing an analysis. Use references/output-template.md when presenting the result.
Workflow
1. Establish the evidence
Collect or infer only what the supplied evidence supports:
- Asset, market, timeframe, date, and data freshness
- Price versus MA20, MA50, and MA200
- MA50/MA200 ordering, crossings, and slopes
- MACD DIF, DEA/signal, histogram direction, zero-axis position, and divergence
- TD direction, setup count, countdown status, timeframe confluence, and confirmation candle
- Bollinger middle band, band position, bandwidth percentile or recent range, squeeze/expansion state, and volume on breakout
- Relevant volume, macro, liquidity, positioning, or on-chain context when available
If current market data is required, retrieve it with an appropriate live-data tool. State the observation timestamp. Never fabricate missing values. If only a chart image is supplied, mark visually estimated levels as approximate.
2. Classify the market regime first
Assign exactly one primary regime and optionally one transition state:
- Bull trend: price above rising MA200; MA50 generally above MA200.
- Bear trend: price below falling MA200; MA50 generally below MA200.
- Range: MA200 flat; MA50/MA200 intertwined; MACD repeatedly crosses near zero.
- Squeeze: Bollinger bandwidth near a medium/long lookback low; direction unresolved.
- Volatility expansion: bandwidth rising after a breakout; direction follows the confirmed break.
- Exhaustion transition: momentum fades and TD/Bollinger evidence warns that the current regime may be ending.
Treat MA200 as an environment filter, not an entry timer. Give its slope more weight than a raw crossing.
3. Read the four layers in order
- Direction — MA50/MA200: decide which side is structurally favored or whether trend methods should be disabled.
- Momentum — MACD: determine whether the favored direction is accelerating, decelerating, or diverging. Prefer weekly evidence over daily, and daily over intraday.
- Exhaustion — TD: treat setup 9 as an alert region, not an automatic reversal. Require countdown/sequence context, higher-timeframe confluence, or a confirmation candle before acting against a move.
- Volatility — Bollinger: decide whether price is compressing, expanding, walking a band, or reverting toward the middle band.
Do not double-count correlated evidence. MA and MACD both derive from price smoothing; volume, liquidity, positioning, fundamentals, or on-chain data add more independent confirmation.
4. Resolve conflicts with hierarchy
Apply this priority:
- Data quality and timeframe
- Macro/liquidity shock or known event risk
- MA200 regime and slope
- Confirmed price breakout/breakdown with volume
- Bollinger volatility transition
- MACD momentum state
- TD exhaustion alert
When signals conflict, lower confidence and preserve both scenarios. Never average contradictions into false certainty.
5. Build a scenario map
Produce three conditional paths when evidence permits:
- Base case: most consistent with the current regime.
- Bull case: trigger, confirmation, targets/next zones, and invalidation.
- Bear case: trigger, confirmation, targets/next zones, and invalidation.
Separate observation from inference. Use language such as “if X closes above Y with expanding bandwidth and volume, then…” Avoid “will,” “guaranteed,” and unsupported probability numbers.
6. Add risk controls
- Distinguish spot, leveraged, and short exposure.
- Size risk from the invalidation level, not conviction.
- Prefer closing-price confirmation over intrabar touches.
- Flag event gaps, liquidation cascades, illiquidity, and manipulated small-cap markets as model-breakers.
- For leveraged trading, emphasize that a correct thesis can still be liquidated before confirmation.
- State that the analysis is educational and not personalized financial advice.
Non-negotiable rules
- Do not reduce the system to “golden cross buy, death cross sell.”
- Do not short merely because price touches the upper Bollinger band or buy merely because it touches the lower band.
- Do not guess the direction of a squeeze; wait for a confirmed break.
- Do not treat TD 9 as an exact reversal timestamp.
- Do not use mean-reversion logic in a strong trend without explicit evidence that the regime changed.
- Do not cite historical win rates unless they were verified for the same asset, timeframe, definitions, and sample.
- Do not optimize parameters on a tiny sample and present them as universal.
Minimum viable analysis
When data is incomplete, still return:
- Known observations
- Provisional regime
- Missing evidence that matters
- Conditional bull and bear triggers
- Invalidation logic and risk note
