| name | tw-stock-trend |
| description | Trend identification for Taiwan-listed stocks — Dow theory structure (HH/HL vs LH/LL), multi-timeframe alignment, strength and phase, pullback vs reversal. Use this skill to answer "is this stock in an uptrend or downtrend, and how strong?" before any entry / exit decision.
|
| category | finance |
| tags | ["stock","taiwan","tw-stock","technical-analysis","charting","indicators"] |
| keywords | ["trend","uptrend","downtrend","Dow theory","higher high","lower low","swing point","ADX","pullback","reversal","break of structure","Wyckoff","趨勢","上漲","下跌","K線","多頭","空頭","回檔"] |
| related | ["tw-stock-technical","tw-stock-chip","tw-stock-fundamental","tw-stock-quant","portfolio-construction"] |
Taiwan Stock Trend Identification
"Is this stock going up or down?" sounds simple, but it's the first question every other decision depends on. Get the trend wrong and even perfect entry timing loses money.
When to Use This Skill
- Deciding whether to look for long or short setups on a specific stock
- Reading TAIEX or sector regime before picking entries
- Distinguishing a pullback (still in trend) from a reversal (trend over)
- Sizing positions: full size in clear trends, small in chop
- Validating that a fundamental or chip-flow thesis isn't fighting price action
Pair with tw-stock-technical for the indicator-level detail (MACD, RSI, KD, Bollinger). This skill is the framework above those indicators.
1. The Definition — Structure, Not Indicators
Indicators lag price. Trend is defined by structure first.
Dow theory in two lines
- Uptrend = Higher Highs and Higher Lows (HH + HL)
- Downtrend = Lower Highs and Lower Lows (LH + LL)
- Sideways = Neither; swings overlap
Uptrend Downtrend Sideways
/\ \ /\ /\
/ \ /\ /\\ / \ / \
/\/ \ / \ / \\ / \/ \
/ \/ \ → / \\ → / \
/ \\
What counts as a "swing high / low"?
-
A swing high is a candle whose high is higher than the N candles to its left and right (typically N=3 on daily, N=5 on weekly). Same with swing lows.
-
Filter by significance. On a noisy stock, use a 2–3% minimum move between swings; otherwise every tiny wiggle counts and the structure becomes meaningless.
-
Mark the last two swing highs and last two swing lows. That's all you need to read structure:
- Last swing high > prior swing high? HH
- Last swing low > prior swing low? HL
- Both true → uptrend confirmed
- Either flips → trend in question
2. Three Timeframes — Dow's Three Trends
| Frame | Horizon | Use |
|---|
| Primary (monthly / weekly K) | Months to years | Strategic direction; long-term position |
| Secondary (weekly / daily K) | Weeks to months | Tactical entry; medium-term swing |
| Minor (daily / hourly K) | Days to weeks | Execution timing; short-term scalp |
Multi-timeframe alignment
| Primary | Secondary | Minor | Action |
|---|
| Up | Up | Up | Full long — strongest setup |
| Up | Up | Down | Buy the pullback — best risk/reward |
| Up | Down | Down | Wait — secondary correction within primary up |
| Down | Up | Up | Counter-trend bounce — small size, fast exit |
| Down | Down | Down | Full short / cash |
| Down | Down | Up | Sell the rally — counter-trend retracement |
-
The larger frame sets direction; the smaller frame sets timing. Don't open a primary-down trade because a minor frame turned up.
-
Conflicts default to the higher frame. Weekly up + daily down → wait, don't fight the weekly.
-
For Taiwan retail with day jobs, the daily K is usually the action frame, weekly K is the trend frame. Hourly is noise unless you trade actively.
3. Indicator Overlay — Confirming Structure
Structure alone is enough to identify a trend, but indicators filter noise and measure strength.
Moving average stack
-
The 20/60/120/240 MA stack confirms trend. Bullish stack (MA5 > MA20 > MA60 > MA120 > MA240) = strong uptrend; bearish stack = strong downtrend. Tangle = no trend, look for range.
-
MA240 as primary bull/bear divider. Price above MA240 with MA240 sloping up = primary uptrend; below with slope down = primary downtrend. Crossing MA240 is a structurally significant event.
-
MA20 slope as secondary signal. Up-sloping MA20 = secondary uptrend; flat = range; down-sloping = secondary downtrend. Slope, not just position.
ADX — Is there a trend at all?
| ADX value | Meaning |
|---|
| < 20 | No trend (range / chop) — trend strategies fail |
| 20–25 | Trend forming |
| 25–40 | Clear trend |
| > 40 | Very strong (and often near exhaustion) |
-
ADX doesn't tell you direction. It tells you whether there's a trend. Pair with structure or MA slope for direction.
-
Below 20 ADX = trade differently. Switch to range strategies (buy support, sell resistance, mean-revert). Most trend-following losses come from forcing trend trades in chop.
4. Trend Strength — How Convicted Is the Trend?
| Strong trend | Weak trend |
|---|
| Shallow pullbacks (3–5% in uptrend) | Deep pullbacks (>10%, near MA60) |
| Pullbacks die at MA20 / MA10 | Pullbacks repeatedly break MA20 |
| Volume rises on impulse, falls on pullback | Volume flat or inverted |
| ADX rising and > 25 | ADX < 20 or falling |
| Few overlapping bars; clean swings | Lots of overlap; choppy candles |
| MAs fan out (widening separation) | MAs converge |
-
Pullback depth measures conviction. A 3% pullback in a $100 stock that holds MA20 is gold; a 12% pullback that pierces MA60 is a different animal.
-
Volume on impulse vs pullback is the most underused signal. Uptrend with rising-volume rallies and falling-volume pullbacks → real demand. The reverse → distribution.
-
A strong trend can stay overbought for weeks. Don't fade RSI > 70 or KD > 80 in a strong trend; that's where the move is. See tw-stock-technical rule on "indicator saturation".
5. Trend Phase — Wyckoff's Four Phases
Trends move through a cycle:
Accumulation → Markup → Distribution → Markdown → Accumulation
(sideways (impulse (sideways (impulse
at bottom, up, at top, down,
quiet) crowd in) crowd selling) crowd out)
How to recognise each
| Phase | Price | Volume | Behaviour |
|---|
| Accumulation | Sideways range after a downtrend | Quiet, then occasional spike on lows that hold | Smart money buying; retail bored |
| Markup | Trending up, HH/HL | Rising on rallies, falling on pullbacks | Retail piles in late |
| Distribution | Sideways range after an uptrend | High but choppy, big up-down candles | Smart money selling into strength |
| Markdown | Trending down, LH/LL | Rising on declines, falling on bounces | Retail holds, hopes, eventually capitulates |
-
Distribution looks like accumulation at the wrong end. The price action is similar (sideways with big bars); the context (after a sustained rally vs after a sustained decline) decides which.
-
Accumulation often coincides with bad fundamentals; distribution often coincides with euphoric news. Pair this with tw-stock-fundamental and tw-stock-chip to disambiguate.
6. Pullback vs Reversal — The Hardest Call
In an uptrend, when price drops, is it a healthy pullback or the start of a downtrend?
Pullback (still in uptrend)
- Holds the most recent HL or bounces from MA20 / MA60
- Reduces depth and volume as it goes
- Resumes within a few bars (typically 3–10 days)
- Makes a new HL above the previous HL on the bounce
Reversal (uptrend probably over)
- Break of structure (BoS): closes below the most recent HL → uptrend invalidated
- The bounce that follows fails to make a new HH (only LH)
- Two failed attempts to break the prior high
- Volume profile flips: heavier on declines than rallies
- MA20 turns down and price stays below
-
The single sharpest signal: HL gets violated. That's "break of structure". Below the most recent HL, the uptrend is no longer technically intact.
-
One BoS is a warning, not a verdict. Wait for a second leg down (LH followed by LL) before declaring reversal. Many "BoS" events are false breaks that snap back.
-
The reverse applies symmetrically. In a downtrend, the most recent LH being broken is the warning; a HH after it confirms reversal.
-
In strong trends, "deep pullbacks" can look like reversals but aren't. That's why depth alone isn't enough — combine with BoS and volume.
7. Range vs Trend — Don't Force a Trend on Chop
The single biggest mistake in trend trading is trading trend strategies in ranges.
Range signals
- ADX < 20
- Price bouncing between two parallel horizontal lines
- MAs tangled (MA5/20/60 within 1–2%)
- Bollinger Bands narrow ("squeeze")
- Volume flat or randomly distributed
-
In a range, switch strategy: buy near support, sell near resistance, mean-revert. Trend signals (MACD cross, golden cross) chronically lose money here.
-
A range often ends in a breakout that becomes the new trend. The transition is hard to see in real time; that's why patient breakout traders wait for the close above the range high (or below the low) with rising volume.
-
TAIEX and individual stocks can be in different regimes. Index ranging while a single stock trends, or vice versa. Read both.
8. Entry Rules — Trading With the Trend
Once you know the trend direction and strength, the rules write themselves:
-
Enter on pullbacks, not breakouts. In an uptrend, the best entries are pullbacks to MA20 / MA60 / a prior breakout level (now support). Chasing breakouts gets you in at peak emotion.
-
Stop below the most recent HL (in an uptrend) or above the most recent LH (in a downtrend). That's the structural invalidation point — beyond it the trade thesis is wrong.
-
Trail stops along subsequent HLs. Each new HL becomes the new stop level. This lets winners run while protecting against giveback.
-
Size proportional to trend strength. Full size in clear primary up + secondary up + minor pullback alignment. Half size when frames disagree. Skip when chop or against primary trend.
-
Never average down in a downtrend. "Cheap" in a downtrend stays cheap, then gets cheaper. Adds only on confirmed HLs in an uptrend.
9. TWSE-Specific Notes
-
Red K = up, Black K = down (opposite of US/European charts). Calibrate when reading translated TA material.
-
10% daily price limit distorts structure analysis. A "limit-up" candle truncates what would be a longer impulse; consecutive limit-ups can mask exhaustion. Look at weekly K for cleaner structure in volatile names.
-
Monthly revenue release (10th of each month, before 23:59) creates predictable jumps — both up and down — that aren't pure technical. Don't read a single post-revenue candle as a structural signal; wait for the second day's confirmation.
-
Ex-dividend dates create technical gaps that aren't real bearish signals. Adjust the chart for dividends (價格還原) or note ex-div days when reading swings. See tw-stock-data on dividend adjustment.
-
First and last 5 minutes of the session are noisy due to opening / closing call auctions. Intraday traders weight 09:05–13:25 mid-session structure higher.
-
TAIEX trend filter dominates individual stocks. When TAIEX is in primary downtrend, ~80% of stocks follow. Lone uptrends exist but are exceptions; default to "wait for TAIEX to turn" unless you have a strong stock-specific thesis.
10. Common Traps
| Trap | Description | Counter |
|---|
| Hindsight trend | "It was clearly up the whole time" — easy to see after the fact | Decide on the last 30 bars visible at the moment |
| Single-candle reversal call | One big bear candle ends an uptrend in your head | Wait for BoS + LH confirmation |
| Trend overfitting | Drawing 3 trendlines on the same chart to find one that fits | One per timeframe, max |
| MA worship | "Price touched MA60, must bounce" | MAs are guides, not laws — combine with structure |
| Counter-trend hero | Repeatedly shorting a strong uptrend on RSI > 70 | Saturated indicators in strong trends are normal |
| Range as trend | Treating 3 weeks of sideways as the start of a trend | Check ADX < 20; use range tactics |
| News override | "Earnings beat, must be uptrend now" | News changes fundamentals; trend needs structural confirmation |
| Stale trend | Calling a 6-month-old uptrend "still up" despite a clear BoS 4 weeks ago | Re-read structure on every entry, not from memory |
11. Pre-Decision Checklist
Before any trade decision, run this top-to-bottom:
If you can't answer all of these, you don't know the trend well enough to size up.
12. Quick Decision Tree
Is ADX > 25?
├── No → range / chop → use range tactics, not trend tactics
└── Yes → there's a trend
│
Is MA240 slope up?
├── Yes → primary uptrend
│ │
│ Is weekly K making HH + HL?
│ ├── Yes → strong primary uptrend → look for long pullbacks
│ └── No → primary up but secondary correction → wait
│
└── No → primary downtrend
│
Is weekly K making LH + LL?
├── Yes → strong primary downtrend → cash / short rallies
└── No → primary down but bouncing → counter-trend, small size
Related Skills
