4 Ways to Read the Heatmap: Price Change Is Just One

4-section structure: Concept / How We Compute / How to Read / Caveats.

1. Concept

Most heatmaps show only price change. But price change only answers "who went up today." It can't answer deeper questions:

"This stock is +3%, but versus the market — strong or weak?"
"Which stocks were most actively traded today?"
"Where did institutional money concentrate?"

Our heatmap now supports 4 modes:

Mode	What It Answers
Price Change	Who gained? Who lost? How much?
Relative Strength (RS)	Who beat the market? Who lagged?
Liquidity	Who is today's actively-traded hot stock?
Amount	Where did money concentrate?

Same heatmap, switch modes, get 4 completely different insights.

2. How We Compute

2.1 Price Change (chg_pct)

Same as Heatmap Time Periods cumulative return. Color scale dynamically adjusts per period (±9.5% ~ ±40%).

2.2 Relative Strength Z-Score

Not absolute return — performance relative to the market:

RS_z = (stock_chg_pct − market_mean_chg_pct) / market_std

Market mean: arithmetic average of all stocks in the index
Market std: standard deviation across the index universe
Unit: σ (standard deviations)

Color scale ±2.0σ:

+2σ deep red: stocks significantly outperforming the market
−2σ deep green: stocks significantly underperforming
0σ neutral: tracks market performance

2.3 Liquidity (Turnover Rate %)

Turnover = daily_volume / shares_outstanding × 100%

Measures "what % of the float changed hands today." Color uses percentile mapping (prevents a few extreme-volume stocks from crushing the scale):

Light orange → deep orange (single-sided, no positive/negative)
Dark = turnover in the top percentile of the market
Light = turnover in the bottom percentile

2.4 Trade Amount (Million NT$)

Amount = daily_volume × close_price

Measures "how much capital flowed through this stock today." Since amount spans 4 orders of magnitude (millions → hundreds of billions), also uses percentile mapping:

Light blue → deep blue
Dark = amount in top percentile (capital concentration)
Light = amount in bottom percentile (ignored)

2.5 Query String Persistence

Mode stored in URL ?mode=rs, composable with market and period:

/home/heatmap?market=上市&period=20d&mode=rs

Switching any one preserves the other two. Shareable links carry full state.

3. How to Read

3.1 Price Change vs Relative Strength: The Common Confusion

Key difference: price change shows absolute, RS shows relative.

Scenario A: Market up day

Index +3%. Stock +2%.
Price change mode: red (+2% is a gain)
RS mode: green (underperforms by 1%)

Scenario B: Market down day

Index −4%. Stock −1%.
Price change mode: green (−1% is a loss)
RS mode: red (outperforms by 3%, fell less than others)

Practical takeaway:

Price change alone is dragged by overall market sentiment
RS helps you find genuine leaders on broad-up days
On broad-down days, helps identify defensive names (may lead the rebound)

3.2 Liquidity × Price Change Combined Reading

Price Change	Liquidity	Meaning
Deep red (up)	Deep orange (active)	Price-volume rising together — strong advance with healthy turnover
Deep red (up)	Light orange (quiet)	Rising on low volume — no sellers, possibly late-stage pump (risk)
Deep green (down)	Deep orange (active)	Panic selling — heavy volume, possibly near capitulation
Deep green (down)	Light orange (quiet)	No-volume decline — no buyers, further downside risk

3.3 Amount Mode: Where Is the Smart Money?

Even among large-caps, amount concentration varies hugely:

Sea of deep blue: capital-concentrated chasing, today's focus stocks
Mid-tone: normal trading
Light blue: capital-avoided, big market cap but no one trades

Read with industry blocks: whole industry block deep blue → sector-wide attention; single stock notably darker → individual event (earnings, news)

3.4 Cycling 4 Modes for a Market Health Check

Standard scan routine:

chg_pct: more gainers or losers today?
rs: who are the true leaders (excluding market drift)?
liquidity: which are the hot active names?
amount: where is the capital concentrated?

Cross-referencing 4 modes is far more complete than price-change alone.

4. Caveats

⚠️ RS Is "Relative," Not "Absolute Good"

RS deep red only means outperforming the market — it doesn't mean going up. On a −5% market day, a stock at −2% shows red RS but is still losing money. Verify absolute direction in chg_pct mode.

⚠️ Market Mean Affected by Extreme Values

RS's "market mean" is the arithmetic average of all stocks. If a cluster of limit-up/limit-down days occurs, the mean tilts.

Mitigate: only rely on significant deviations (±1.5σ+), ignore tiny ±0.3σ differences.

⚠️ Turnover Is Intraday, Not Cumulative

Liquidity mode shows today's turnover rate; it doesn't change with period. "Cumulative turnover" isn't provided (requires additional aggregation cost).

⚠️ Amount Percentile Changes Per Index

Switching market (上市 / 上櫃 / 興櫃) recomputes the amount percentile baseline. So the same stock's color depth may differ between markets — rank depends on the sample pool.

⚠️ Emerging Market Turnover Spikes

Emerging-market (興櫃) stocks have low floats and thin volume, so turnover easily spikes to 10%+. Liquidity color in 興櫃 clusters at the low end (a few extreme stocks eat the visual scale). This is a structural limitation, not a bug.

⚠️ Excludes After-Hours and Odd-Lot Trading

Amount and liquidity only count regular-session matched trades, excluding after-hours fixed price and odd-lot trading. Large-caps' true volume is slightly understated (after-hours may add 10-20%).

Try It

Open Heatmap, cycle 4 modes in the header
Toggle between price change and RS, watch which stocks "flip red to green" or vice versa
Switch to mode=amount to see where capital concentrated among large-caps
Combine period=20d + mode=rs → find 20-day market-beaters
Click 📐 for RS formula and percentile-mapping rules