07 · Drawing Tools in Practice

Drawing is the only way to translate "the price structure you understand" into "a verifiable hypothesis on the screen". Horizontal lines mark support/resistance, trend lines mark rhythm, channels mark boundaries, Fibonacci marks retracement targets, XABCD marks harmonic structure — the tools themselves have no predictive power; the quality of a drawing depends on "why you drew it here", not "how precisely it is drawn".

💡 One-Sentence Summary

One-sentence summary: drawing exists to give trading decisions a "frame of reference" and "invalidation conditions", not to produce pretty shapes. A line without a corresponding stop-loss level or observation level is decoration.

1. The Essence of Drawing: Translating Structure into Hypotheses

Three classes of drawing tools: positioning · connection · proportion

There are hundreds of drawing tools in technical analysis, but every tool essentially does one of three things:

Tool class	Question it answers	Typical tools
Positioning	"Where are the important price levels/times?"	Horizontal line, vertical line, price label
Connection	"What is the rhythm/boundary/trend of this move?"	Trend line, ray, channel, speed lines
Proportion	"Where are the retracement/extension/pattern targets?"	Fibonacci family, XABCD, Gann

The three levels of correct usage:

Describe (lowest): draw the structure that already happened — anyone can do this, but it has only "hindsight verification" value;
Presuppose (middle): have a view before drawing ("this is resistance at the prior high"); the line is just the visualization of that view;
Trigger (highest): the line gives explicit trigger conditions — breakout/breakdown/retest confirmation, executed together with the risk control of the 07 · Trading Systems chapter.

Interaction conventions (common to all tools; not repeated below): every anchor is draggable for fine-tuning; a selected drawing can be dragged as a whole; hit-testing differs by shape — segment-type tools hit by distance to the segment, band/channel/range-type tools hit by area (any point inside the band is clickable), and once hit the object is selected and editable. Layer organization (hide/lock/delete) is covered in 5.10.

⚠️ Red Line

The vast majority of beginners die from "drawing first, finding reasons after". The correct order is always: form a hypothesis first → mark its key levels with drawings → wait for price to verify or falsify it.

2. Positioning: Support/Resistance and Key Levels

2.1 Horizontal Line

The horizontal line is the single most important tool; what it draws is not "a line" but a price anchor:

Where to draw: prior highs/lows, round-number gates, edges of high-volume nodes (with the volume profile from 03 · Volume-Price Analysis), near long-term moving averages;
How to use: the first touch is usually just a "test"; the second/third touch is where the real battle happens — the more touches and the longer price lingers, the more "gold" the line carries;
Invalidation: after an effective close-through break (usually 1.5%–3% or 1 ATR; see the ATR section of 02 · Technical Indicators in Depth), support becomes resistance and resistance becomes support.

2.2 Price Label and Vertical Line

Price label: pins "the level I currently consider important" directly on the chart. Label only 3–5; labeling everything equals labeling nothing;
Vertical line: marks time, not price — earnings days, CPI releases, delivery dates, large unlock dates. For event-driven usage see the event-driven section of 11 · Trading Practice.

3. Connection: Trend and Rhythm

3.1 Trend Line

A trend line connects two or more significant lows (uptrend) or highs (downtrend):

Minimum two points, the third confirms: two points make a line but only a "guess"; a third touch that holds is what earns respect;
The meaning of slope: the steeper the slope, the faster it tends to fail (steep trend lines, once broken, usually lead to outright reversal rather than sideways); the gentler the slope, the longer the reference value lasts;
Drawing principle: connect body edges (wick piercings are normal — don't force connections to wick extremes), and prefer lines "tested multiple times".

3.2 Ray and Polyline

Ray: extends infinitely to the right from its origin — fits "start from a historical key level and project the future extension direction", e.g., drawing a ray from a major bottom to see the current trend's extension boundary;
Polyline: manually "straightens" a complex move into segments for structural review (which leg was the markup, which the correction); not suitable as a direct trading trigger, because you chose the anchors yourself — inherently subjective.

3.3 Arrow

The arrow is "annotation", not "analysis": mark on the chart "this breakout was valid" or "this was a false breakout". Best used with the screenshot-sharing feature — during review, arrows + text reconstruct what you were thinking at the time; see the advanced trade-review section of 07 · Trading Systems.

4. Channels: Boundaries and Regression

4.1 Parallel Channel and Horizontal Channel

Parallel channel: one trend line + one parallel boundary line framing a trend's "upper and lower rails". Usage:
- Buy the lower rail, sell the upper rail, stand aside at the middle (oscillating channel);
- A channel break ≠ chase immediately: after a break, a "retest confirmation" before continuation is common; chasing directly gets harvested by false breakouts;
Horizontal channel: two horizontal lines framing a sideways range (a box). The box's value lives on its edges: buy the lower edge, sell the upper edge; after the box breaks, the old box becomes a "support/resistance shelf".

4.2 Regression Channel and Speed Lines

Regression channel: runs a linear regression over all candles in the range and draws "best-fit midline ± standard deviation" rails. More objective than hand-drawn channels — it doesn't depend on your shaky anchor choices; well suited to verify "is the current trend still running on one side of the midline";
Speed lines: divide the A→B advance into thirds and draw 1/3 and 2/3 speed lines. Use: in a trend, a pullback holding the 1/3 line is usually strong consolidation; losing the 2/3 line often means the trend is accelerating downward. A "secondary reference" — mediocre signal quality on its own.

5. Proportion: The Fibonacci Family

5.1 Fibonacci Retracement

The most-used item: from one significant swing (e.g., low → high), pull the 0.236/0.382/0.5/0.618/0.786 retracement levels:

0.618 is the "golden level": a pullback to 0.618 that holds → healthy trend; a close below the 0.618 retracement → the original trend most likely weakens;
Usage: at retracement levels (especially near 0.382/0.5/0.618) wait for stabilization signals (e.g., hammer/engulfing from 01 · Candlestick Patterns + expansion) — do not place orders directly at the levels;
Common mistake: treating retracement levels as "exact points" instead of "fuzzy zones". ±1% around 0.618 all counts as valid; chasing two decimal places of precision is self-deception.

5.2 Fibonacci Extension / Fan / Time Zones

Extension (Fib Extension): projects post-breakout targets (1.272/1.618/2.618). Use: a take-profit reference, not "a price that must be reached" — near the target, scale out and trail the stop up;
Fan (Fib Fan): radiates ratio lines of different slopes from the origin, used to judge the "angle class" of a trend pullback;
Time zones (Fib Time): draws vertical lines at golden-ratio divisions of the time axis to project "regime-change time windows". Honest assessment: time tools' win rate is close to random — better used as "raise alertness when the time point arrives" reminders than as trading signals.

5.3 Cycle Lines

Cycle lines take the first anchor A as origin, treat the A→B interval as one cycle, and extend 12 equidistant vertical lines to the right (a solid line at A, then dashed lines with +1/+2… sequence labels). Good for:

Watching whether the rhythm "recurs": start A at an obvious swing low, measure to the next similar low B; the extended lines land on later "potential in-phase turning points";
Cross-verification with Gann/Fibonacci time lines: different time tools resonating on the same day deserve extra alertness;
Detecting cycle drift: if turns keep landing "a few days late" versus the extensions, the cycle is lengthening — don't mechanically bottom-fish on the original cycle.

5.4 Fib Channel

The Fib channel = one base trend line + a set of channel lines offset in parallel by Fibonacci ratios. After dragging A→B to define the swing (base line), the system treats the A→B swing as 1× and generates 8 parallel ratio lines at 0.236/0.382/0.5/0.618/0.786/1/1.272/1.618 on both sides, spanning the full visible width:

Direction-sensitive: when A→B swings up, the ratio lines extend proportionally above the base line (pullback/extension targets); a downward swing mirrors them below. Clicking A first then B sets the channel's orientation — consistent with the "start and end set direction" logic of Fibonacci retracement;
Key ratios: 0.382/0.618 are the pivots "most tested by retests inside the channel"; 1.272/1.618 become the extension target band after the channel breaks (like the extension tool); 0.236 is the watch level for "shallow pullbacks" in strong trends;
Usage: while price runs inside the channel, the lower/upper rails are zone boundaries, not exact points; after an effective break, the old boundary flips into support/resistance and the extension levels (1.272/1.618) are references for scaling out;
Difference from the parallel channel: the parallel channel draws only equal-width rails (assuming constant swing size), while the Fib channel maps the golden ratios of the same swing into multiple parallel lines — better for "after one markup wave, judging pullback depth and extension targets".

Honest assessment: channel spacing is set by the A→B swing size — the longer the anchors, the wider the channel and the lower the reference value (it can wrap almost any market). Draw on "structurally clean swings" first, not force-fit onto chop.

5.5 Wedge

The wedge uses three converging edges to describe the "compression" late in a trend: points 1/2 are the origins of the two edges, point 3 is the convergence point, and the two edges continue as dashed projections beyond it. Two kinds by direction:

Rising wedge: both edges slope up and converge; common in late-stage advances — bull strength decaying, new highs on flattening slopes; losing the lower edge is a bearish signal;
Falling wedge: both edges slope down and converge; common in late-stage declines — bear strength exhausting; clearing the upper edge is a bullish signal.

Usage:

Don't act inside the wedge: convergence means the range is narrowing and direction is undecided before the break — blind fade trades get slapped;
Trade the breakout direction: a rising wedge losing its lower edge (with expansion) is bearish; a falling wedge clearing its upper edge (with expansion) is bullish; false breakouts are common at wedge ends — entering after a retest confirmation is steadier;
Meaning of the convergence point: the closer the edges to the apex, the nearer the regime change — treat the area around the apex as a "time warning zone";
Difference from triangle/channel: a triangle is a "converging pattern"; a wedge is a "directionally tilted triangle"; channel edges are parallel (constant swing size) while wedge edges converge (shrinking swing size).

Honest assessment: the wedge is one of the most abused shapes — in chop, a casual drag produces "wedges" everywhere. Use it only in structurally clean late-trend phases (a clear prior impulse + visibly slowing slope now), and always give the wedge a breakout-confirmation invalidation condition (e.g., an effective close-through + 1 ATR).

5.6 Parallel Ray / Width Channel

These two tools serve the same need: after drawing one trend line, quickly replicate "same slope, different location" reference lines.

Parallel ray: click A, B to define direction, then C as the ray origin — the system extends a ray from C, exactly parallel to A→B, infinitely to the right (A→B shown as a thin dashed direction line). Typical use: "carry" the slope of a main trend to another price level and watch whether same-slope, different-level action recurs (e.g., shifting the prior-high trend line down to the current low as a support reference);
Width channel: click A, B to define the base direction, then C to set the channel width — the system draws two infinite parallel lines through A and through C, with a dashed line marking the B→C width. Difference from "parallel channel": the parallel channel's width is fixed by the A→B price gap, while the width channel's width is freely set by the third point C — better when "each edge should hug a separate swing".

Usage:

Parallel ray = slope transfer: define the slope on a clear impulse A→B, then shift it to the key level you care about (prior high/low, round-number gate). Do not keep re-transferring a slope in markets where it has already failed — trend-line slope itself changes across markup/markdown legs;
Width channel = edge hugging: C should sit on "the high/low of another real swing", letting the rails hug real action — not a casually dragged width; the channel's reference value depends entirely on whether C hugs real structure;

Honest assessment: parallel tools are natural extensions of the "trend line" and generate no new information — they only replicate the slope of existing structure. The genuinely effective use: A→B defines a large-degree reliable trend, C lands on a small-degree key level, forming a "big slope + small location" combined observation — not doodling two parallel lines on a 5-minute chart for self-amusement.

5.7 Trend Angle / Time Range

These two tools make "slope" and "time window" explicit — auxiliary observation drawings: they add no structure, but annotate quantitative information you could already see but tend to overlook.

Trend angle: drag A→B to draw a segment; the system draws a small arc at the A end and labels the angle relative to horizontal at the midpoint (screen space; positive upward, negative downward, range −90°~+90°). The two points are time-ordered (left first), and once selected the whole line can be dragged or anchors fine-tuned;
Time range: drag A→B to frame two time points; the system draws a semi-transparent vertical band on the main chart (left/right borders + a top date-range label "start ~ end"). Anywhere inside the band is clickable — good for marking "event windows", "earnings/data release sessions", or "the time span of a completed structure".

Usage:

Angle for comparing slope changes: the angle difference between markup legs and pullback legs quantifies quickly — e.g., "this rally runs at 38° versus 52° for the last one, a clear slowdown" — more objective than eyeballing two lines. Note: the chart's horizontal axis (time) and vertical axis (price) have different dimensions; the angle is a screen-space value that changes with chart width/zoom — compare only within the same view;
Time range as "time anchors" for the chart: make the borders land on key time points (pattern start/end, major news moments), not a casual drag — a range's value lies in "what happened during this window".

5.8 Price Band

The price band is the horizontal twin of the time range: the time range paints "a period" as a vertical band; the price band paints "a price span" as a horizontal band across the whole chart.

Dragging: drag A→B (vertical span); the system normalizes the two points by price order (lower price on top edge, higher price on bottom edge) and draws a semi-transparent horizontal band — two border lines spanning the full width, each with a price label on the left (toFixed(2));
Typical use: marking high-volume zones / support-resistance bands / target-price zones — e.g., "64_000~64_300 below is the last two weeks' high-volume band; losing the lower edge means the bulls' line has failed". It expresses "a zone" better than a single horizontal line, and fits "care only about price, not the time window" better than a rectangle.

Honest assessment: the price band is the zonal expression of support/resistance and predicts nothing. Its value is correcting the "one line misread as an exact value" trap into "this is a fuzzy zone" — price weaving inside the band is normal; what carries signal meaning is the break/loss of the band edges (read together with volume and the volume profile).

5.9 Fib Time Zones

Fib time zones are the Fibonacci version of cycle lines: drag A→B to set the base period (A as origin, the A→B interval as the base); the system draws boundary lines to the right of A at Fibonacci multiples (1, 2, 3, 5, 8, 13, 21, 34, 55…) and fills alternating semi-transparent vertical bands between adjacent boundaries — laying "Fibonacci days/weeks" time windows intuitively on the chart.

Dragging: drag A→B (horizontal span); A is the origin and B only fixes the base length; boundaries fall at 1×, 2×, 3×, 5×, 8×, 13×, 21×, 34×, 55× the base from A (the first n=1 line solid, the rest dashed), with multiple labels on top;
Direction-sensitive: like cycle lines, the two points are order-preserving — A on the left extends right, A on the right extends left; flipping the drag direction mirrors the zones and won't auto-normalize away your intent;
Typical use: marking Fibonacci time windows — set A at the start of a significant trend and the base as "one complete swing"; the subsequent boundaries are the theoretical time nodes prone to regime changes/turns. Good for reviewing "did the last few highs/lows land near boundaries" and for pre-marking "where the next time window is" while a trend runs.

5.10 Layer Management: Hide, Lock, and Clear

As drawings accumulate, the chart fills with "old stale lines" — not every line deserves deletion, but not every line deserves to be always visible and clickable either. Layer management answers "how to organize once lines pile up": each line can be hidden (not rendered, not hit-testable) or locked (still rendered, but not selectable or draggable), and can be deleted individually or cleared in one shot.

Hide (👁/🚫): a hidden line disappears from the chart and no longer participates in any hit testing — it is merely "archived"; the stored data remains and it can be shown again anytime. Suits lines "not needed now but needed for later review" (e.g., the multiple parallel channels of some historical swing);
Lock (🔓/🔒): a locked line still renders (the structure stays on chart as reference) but can't be clicked or dragged — mouse/touch on it selects nothing, so no accidental drags. Suits key levels "already confirmed, reference only" (e.g., weekly-degree trend lines), protecting them from accidental bumps during daily zoom/pan;
Per-row delete (🗑): deletes just that one; clear all wipes every drawing of the current pair. Deletion is irreversible — think before clearing;
Row selection state: clicking a row in the panel selects the corresponding drawing on chart (blue + anchors), making it easy to locate a line from the list;
Entry: a "Layer Management" button (with current count) inside the "Drawings" collapsible panel on desktop/mobile; click to open the list.

Honest assessment: layer management is a pure organization tool that generates no analytical signal — what it prevents is "too many lines polluting the view" and "accidentally dragging key levels". Hidden ≠ deleted (data remains, recoverable anytime); locked ≠ invisible (structure remains, just not operable). Sensible usage: lock key levels, hide process lines freely, delete dead lines promptly; leave the chart face to the structures actually in use, and delegate the rest to the layer list.

5.11 Risk/Reward (R:R)

Risk/reward is the most "trader-flavored" auxiliary tool: it lays the entry, stop, and target prices on the chart at once and auto-computes the payoff ratio — an intuitive yardstick for position sizing and "is this trade worth taking".

Drawing: click three points in sequence — A entry → B stop → C target (the order is the semantics; the system preserves order and never re-sorts); three horizontal lines spanning the full width appear: entry solid, stop/target dashed (all turn blue when selected), each with a price label on the left, and a payoff-ratio label (1:{ratio}) near the entry line on the right;
Algorithm: risk = |A − B| (entry to stop distance), reward = |A − C| (entry to target distance), ratio = reward / risk; when the stop equals the entry (risk=0), ratio records 0 without crashing — also a reminder that "a trade without a stop has no payoff ratio to speak of";
Typical use: before placing the order, lay out on the chart "where I exit if wrong, where I exit if right" and see at a glance whether R:R ≥ 2 (usually at least 1:2 to justify the risk); also good for review — losing trades are often not wrong on direction but on an inverted payoff ratio (taking trades where reward is far smaller than risk).

Honest assessment: R:R is a risk-control-first tool, not a prediction tool — it doesn't tell you "will it rise"; it puts "how much if wrong, how much if right" on the table. A pretty payoff ratio still loses if the win rate is too low; real position decisions combine R:R with the win rate to compute expectancy (EV = win rate × reward − loss rate × risk). No stop, or a carelessly tight stop (swept by normal volatility), are the most common misuses of the R:R tool.

6. Shapes: Rectangle / Ellipse / Circle / Triangle / Arc

Shape tools are auxiliaries that "turn what the eye sees into a measurable object":

Tool	Use	Discipline
Rectangle	Frame a range/consolidation platform	Only the edges matter; don't trade inside
Triangle	Frame a converging/diverging pattern	The breakout direction matters more than the shape
Ellipse/circle/arc	Mark arc structures (rounding bottom/top)	Most subjective; review annotation only

Core reminder: the "shape" drawn by these tools only has operational value together with the confirmation signals of 04 · Advanced Candlesticks (breakout + volume + retest). A bare triangle outline says nothing.

7. Structure: XABCD Harmonics / Elliott Waves / Gann Angles

7.1 XABCD Harmonic Pattern

Harmonic patterns use five points X→A→B→C→D to describe price structures with specific ratios (Butterfly, Bat, Crab, Gartley, etc.); the core is the Fibonacci ratio relationships between legs:

Usage: find patterns where XA, AB, BC, CD satisfy the classic ratios (e.g., AB ≈ 0.618 of XA, BC ≈ 0.382–0.886 of AB, CD ≈ 1.272–1.618 of BC), then look for reversal signals near D;
Controversy and traps:
- All five points are adjustable in hindsight; win rates are inflated in samples with "pattern first, price after";
- Always wait at D for "reversal confirmation" (reversal candle + volume); never counter-trend bottom-fish or top-pick just because "the ratios arrived";
Recommendation: treat it as a detector of "potential reversal zones", not a precise entry trigger.

7.2 Elliott Wave

The full mechanics of the five-wave structure (1–5 impulse + A–C correction), the three iron rules, and the "thousand people, thousand counts" controversy are covered in 05 · Elliott · Gann · Chan. Practical drawing points:

Use the 5-point tool to mark the ends of waves 1–5 in order — write down your wave-count hypothesis before drawing;
Iron-rule check: wave 2 doesn't break wave 1's start, wave 3 isn't the shortest, wave 4 doesn't overlap wave 1 — any violation means recount;
The right mindset: wave counting is a "descriptive framework"; the tool's value is forcing you to think clearly about "where we are", not handing you deterministic buy/sell points.

7.3 Gann Fan

Nine angle lines from 1×8 to 8×1 out of an origin, representing the "time × price" speed scale:

The 1×1 line (45°) is the key reference: price above it → bullish strength; below → weakening;
Controversy: the angles depend on axis scaling (stretching the chart wider/taller changes all the angles), so Gann lines "morph" across software and zoom levels — always use them within the same coordinate frame, as a trend-speed reference only, never as precise signals.

7.4 Gann Box

The Gann box is the "zonal" version of the fan: drag A→B to frame a rectangle, and 10 angle lines are auto-drawn inside — 1×1 double diagonals (bold) + 1×2 / 2×1 from the four corners.

Usage: after framing a consolidation range, watch price's "angle rhythm" inside the box — riding 1×1 = equilibrium; losing it and switching to 2×1 support = deceleration; steep climb along 2×1 = acceleration leg (beware the emotional climax);
Resonance: spots where the box's 1×1 diagonal / upper edge overlaps Fibonacci 0.5/0.618 or channel boundaries carry more reference value;
For the full practice (including time-price balance and common mistakes), see 08 · Gann Box and Angles in Practice.

8. Annotation: Measure and Text

8.1 Measure

The measure tool shows the price gap and percentage change between two points; best for:

Measuring "breakout height": from the box top to the target, for estimating the equal-move objective after a break;
Measuring "pullback depth": quickly judging whether the current pullback is 0.382 or 0.5, cross-checked with Fibonacci;
Measure gives numbers, not opinions — numbers only mean something inside your trading plan.

8.2 Text

Text annotation is the number-one productivity tool for review: write on the chart "why I bought here", "what I overlooked then", "which link of this trade lost money". Keep a "chart + words" double record for every trade; reviewing a month later exposes your decision holes better than any indicator.

Kline Buty's text annotation supports multi-line text (Enter for newline / ⌘+Enter to confirm), with font size (A−/A+) and color adjustable in the edit panel (theme yellow by default, or blue/red/green/white/purple) — use colors to grade annotations (e.g., red = risk, green = opportunity), so one glance during review separates "plans" from "warnings". The whole text box is click-selectable (not just the anchor), easy to revise or delete later.

9. The Discipline of Drawing: From "Can Draw" to "Can Use"

9.1 Three Iron Rules for Anchor Selection

Anchors must come from significant structure: prior highs/lows, volume-heavy tops/bottoms, long-consolidation boundaries — never random swing extremes;
Fewer anchors, more credibility: a trend line needs 3+ tests to have value; frequent redraws mean you are "accommodating the market";
Multi-timeframe alignment: a line drawn on the daily should still "mean something" on 4H/1H — big-timeframe anchors stay valid on smaller ones, not vice versa (method in 04 · Advanced Candlesticks).

9.2 Every Line Must Have an "Invalidation Condition"

After drawing any line, ask yourself three questions:

Question	Consequence of having no answer
What view does this line represent?	It's decoration — delete it
What do I do when price hits the line?	No trigger plan = not drawn at all
What price action proves this line wrong?	No invalidation condition = you will hold losers forever

💡 The Ultimate Test of Drawing

The ultimate test of drawing: cover all the lines on the chart; if you can still say "where is support, where is resistance, where is my stop, what counts as wrong", the lines are valid; if you can't, they are psychological comfort.

9.3 Common Mistakes Checklist

❌ Over-drawing (more than ~10 lines per screen): information overload equals no information;
❌ Repeatedly moving anchors to "fit" the market: this is drawing-tool overfitting — the same disease as the backtest overfitting covered in 06 · Critique and Validation;
❌ Treating retracement/target levels as exact prices: they are zones, not points;
❌ Replacing risk control with drawing signals: however good the structural call, the stops/position rules of the 07 · Trading Systems chapter are the backstop.

9.4 Mobile & Touch Interaction Differences

Touch has no hover state, so the interaction mindset differs from desktop — four differences worth knowing:

Two-step editing: desktop separates select vs. edit via "hover + click"; touch switches to "tap to select → drag to edit" to avoid accidental touches;
Auto-return to "mouse" after each drawing: on mobile, every completed drawing (including text confirmation) switches the tool back to "mouse" — right after drawing, the finger most likely wants to fine-tune, not draw another; leaving the tool active invites accidental creation; desktop keeps continuous drawing;
Double-click / long-press to rewrite text: desktop double-clicks the text body, mobile long-presses it (~250ms) to open the editor directly — no panel needed; works only in "mouse" read-only mode, so it never clashes with drawing gestures;
Reading data in comfort: after dragging or long-press pinning, the crosshair lingers about 2 seconds on release (a tap or any new gesture clears it at once), and the OHLC tooltip flips at container edges to stay fully inside the viewport — even drags in the mid-lower chart can read the last row.

10. Summary: The "Dao" and "Shu" of Drawing

Level	Content	Corresponding tools
Shu (how to draw)	Anchors from significant structure, multi-point confirmation, multi-timeframe alignment	Trend lines/channels/Fibonacci
Fa (how to use)	Hypothesis first, zones not points, explicit invalidation conditions	Horizontal lines/XABCD/measure
Dao (why)	Drawing is the visualization of thinking, not a prediction grail	All tools

Read this chapter's discipline together with 07 · Trading Systems and 08 · Pitfalls; only then does drawing graduate from "decoration" to "decision".

Conventions

All prices and figures in this article are illustrative and represent no specific instrument or market.
The exact interactions of each drawing tool (anchor count, drag method) are governed by the actual Kline Buty interface.
Fibonacci ratios and harmonic-pattern ratios are "statistical tendencies", not laws; before use, understand the validation methodology of 06 · Critique and Validation.
For counter-trend trades (e.g., harmonic D-point bottom-fishing), be sure to read the margin and liquidation sections of the 03 · Futures chapter first.

⚠️ Risk Warning

Drawing is a subjective tool: two people can draw exactly opposite trend lines on the same chart. Its value lies in disciplining your plan, not predicting the market — however beautiful a counter-trend bottom-fish looks, compute margin and liquidation price first.

Drawing is the only way to translate "the price structure you understand" into "a verifiable hypothesis on the screen". Horizontal lines mark support/resistance, trend lines mark rhythm, channels mark boundaries, Fibonacci marks retracement targets, XABCD marks harmonic structure — the tools themselves have no predictive power; the quality of a drawing depends on "why you drew it here", not "how precisely it is drawn".

💡 One-Sentence Summary

1. The Essence of Drawing: Translating Structure into Hypotheses

Three classes of drawing tools: positioning · connection · proportion

There are hundreds of drawing tools in technical analysis, but every tool essentially does one of three things:

Tool class	Question it answers	Typical tools
Positioning	"Where are the important price levels/times?"	Horizontal line, vertical line, price label
Connection	"What is the rhythm/boundary/trend of this move?"	Trend line, ray, channel, speed lines
Proportion	"Where are the retracement/extension/pattern targets?"	Fibonacci family, XABCD, Gann

The three levels of correct usage:

Describe (lowest): draw the structure that already happened — anyone can do this, but it has only "hindsight verification" value;
Presuppose (middle): have a view before drawing ("this is resistance at the prior high"); the line is just the visualization of that view;
Trigger (highest): the line gives explicit trigger conditions — breakout/breakdown/retest confirmation, executed together with the risk control of the 07 · Trading Systems chapter.

⚠️ Red Line

2. Positioning: Support/Resistance and Key Levels

2.1 Horizontal Line

The horizontal line is the single most important tool; what it draws is not "a line" but a price anchor:

Where to draw: prior highs/lows, round-number gates, edges of high-volume nodes (with the volume profile from 03 · Volume-Price Analysis), near long-term moving averages;
How to use: the first touch is usually just a "test"; the second/third touch is where the real battle happens — the more touches and the longer price lingers, the more "gold" the line carries;
Invalidation: after an effective close-through break (usually 1.5%–3% or 1 ATR; see the ATR section of 02 · Technical Indicators in Depth), support becomes resistance and resistance becomes support.

2.2 Price Label and Vertical Line

Price label: pins "the level I currently consider important" directly on the chart. Label only 3–5; labeling everything equals labeling nothing;
Vertical line: marks time, not price — earnings days, CPI releases, delivery dates, large unlock dates. For event-driven usage see the event-driven section of 11 · Trading Practice.

3. Connection: Trend and Rhythm

3.1 Trend Line

A trend line connects two or more significant lows (uptrend) or highs (downtrend):

Minimum two points, the third confirms: two points make a line but only a "guess"; a third touch that holds is what earns respect;
The meaning of slope: the steeper the slope, the faster it tends to fail (steep trend lines, once broken, usually lead to outright reversal rather than sideways); the gentler the slope, the longer the reference value lasts;
Drawing principle: connect body edges (wick piercings are normal — don't force connections to wick extremes), and prefer lines "tested multiple times".

3.2 Ray and Polyline

Ray: extends infinitely to the right from its origin — fits "start from a historical key level and project the future extension direction", e.g., drawing a ray from a major bottom to see the current trend's extension boundary;
Polyline: manually "straightens" a complex move into segments for structural review (which leg was the markup, which the correction); not suitable as a direct trading trigger, because you chose the anchors yourself — inherently subjective.

3.3 Arrow

4. Channels: Boundaries and Regression

4.1 Parallel Channel and Horizontal Channel

Parallel channel: one trend line + one parallel boundary line framing a trend's "upper and lower rails". Usage:
- Buy the lower rail, sell the upper rail, stand aside at the middle (oscillating channel);
- A channel break ≠ chase immediately: after a break, a "retest confirmation" before continuation is common; chasing directly gets harvested by false breakouts;
Horizontal channel: two horizontal lines framing a sideways range (a box). The box's value lives on its edges: buy the lower edge, sell the upper edge; after the box breaks, the old box becomes a "support/resistance shelf".

4.2 Regression Channel and Speed Lines

Regression channel: runs a linear regression over all candles in the range and draws "best-fit midline ± standard deviation" rails. More objective than hand-drawn channels — it doesn't depend on your shaky anchor choices; well suited to verify "is the current trend still running on one side of the midline";
Speed lines: divide the A→B advance into thirds and draw 1/3 and 2/3 speed lines. Use: in a trend, a pullback holding the 1/3 line is usually strong consolidation; losing the 2/3 line often means the trend is accelerating downward. A "secondary reference" — mediocre signal quality on its own.

5. Proportion: The Fibonacci Family

5.1 Fibonacci Retracement

The most-used item: from one significant swing (e.g., low → high), pull the 0.236/0.382/0.5/0.618/0.786 retracement levels:

0.618 is the "golden level": a pullback to 0.618 that holds → healthy trend; a close below the 0.618 retracement → the original trend most likely weakens;
Usage: at retracement levels (especially near 0.382/0.5/0.618) wait for stabilization signals (e.g., hammer/engulfing from 01 · Candlestick Patterns + expansion) — do not place orders directly at the levels;
Common mistake: treating retracement levels as "exact points" instead of "fuzzy zones". ±1% around 0.618 all counts as valid; chasing two decimal places of precision is self-deception.

5.2 Fibonacci Extension / Fan / Time Zones

Extension (Fib Extension): projects post-breakout targets (1.272/1.618/2.618). Use: a take-profit reference, not "a price that must be reached" — near the target, scale out and trail the stop up;
Fan (Fib Fan): radiates ratio lines of different slopes from the origin, used to judge the "angle class" of a trend pullback;
Time zones (Fib Time): draws vertical lines at golden-ratio divisions of the time axis to project "regime-change time windows". Honest assessment: time tools' win rate is close to random — better used as "raise alertness when the time point arrives" reminders than as trading signals.

5.3 Cycle Lines

Watching whether the rhythm "recurs": start A at an obvious swing low, measure to the next similar low B; the extended lines land on later "potential in-phase turning points";
Cross-verification with Gann/Fibonacci time lines: different time tools resonating on the same day deserve extra alertness;
Detecting cycle drift: if turns keep landing "a few days late" versus the extensions, the cycle is lengthening — don't mechanically bottom-fish on the original cycle.

5.4 Fib Channel

Direction-sensitive: when A→B swings up, the ratio lines extend proportionally above the base line (pullback/extension targets); a downward swing mirrors them below. Clicking A first then B sets the channel's orientation — consistent with the "start and end set direction" logic of Fibonacci retracement;
Key ratios: 0.382/0.618 are the pivots "most tested by retests inside the channel"; 1.272/1.618 become the extension target band after the channel breaks (like the extension tool); 0.236 is the watch level for "shallow pullbacks" in strong trends;
Usage: while price runs inside the channel, the lower/upper rails are zone boundaries, not exact points; after an effective break, the old boundary flips into support/resistance and the extension levels (1.272/1.618) are references for scaling out;
Difference from the parallel channel: the parallel channel draws only equal-width rails (assuming constant swing size), while the Fib channel maps the golden ratios of the same swing into multiple parallel lines — better for "after one markup wave, judging pullback depth and extension targets".

5.5 Wedge

Rising wedge: both edges slope up and converge; common in late-stage advances — bull strength decaying, new highs on flattening slopes; losing the lower edge is a bearish signal;
Falling wedge: both edges slope down and converge; common in late-stage declines — bear strength exhausting; clearing the upper edge is a bullish signal.

Usage:

Don't act inside the wedge: convergence means the range is narrowing and direction is undecided before the break — blind fade trades get slapped;
Trade the breakout direction: a rising wedge losing its lower edge (with expansion) is bearish; a falling wedge clearing its upper edge (with expansion) is bullish; false breakouts are common at wedge ends — entering after a retest confirmation is steadier;
Meaning of the convergence point: the closer the edges to the apex, the nearer the regime change — treat the area around the apex as a "time warning zone";
Difference from triangle/channel: a triangle is a "converging pattern"; a wedge is a "directionally tilted triangle"; channel edges are parallel (constant swing size) while wedge edges converge (shrinking swing size).

5.6 Parallel Ray / Width Channel

These two tools serve the same need: after drawing one trend line, quickly replicate "same slope, different location" reference lines.

Parallel ray: click A, B to define direction, then C as the ray origin — the system extends a ray from C, exactly parallel to A→B, infinitely to the right (A→B shown as a thin dashed direction line). Typical use: "carry" the slope of a main trend to another price level and watch whether same-slope, different-level action recurs (e.g., shifting the prior-high trend line down to the current low as a support reference);
Width channel: click A, B to define the base direction, then C to set the channel width — the system draws two infinite parallel lines through A and through C, with a dashed line marking the B→C width. Difference from "parallel channel": the parallel channel's width is fixed by the A→B price gap, while the width channel's width is freely set by the third point C — better when "each edge should hug a separate swing".

Usage:

Parallel ray = slope transfer: define the slope on a clear impulse A→B, then shift it to the key level you care about (prior high/low, round-number gate). Do not keep re-transferring a slope in markets where it has already failed — trend-line slope itself changes across markup/markdown legs;
Width channel = edge hugging: C should sit on "the high/low of another real swing", letting the rails hug real action — not a casually dragged width; the channel's reference value depends entirely on whether C hugs real structure;

5.7 Trend Angle / Time Range

Trend angle: drag A→B to draw a segment; the system draws a small arc at the A end and labels the angle relative to horizontal at the midpoint (screen space; positive upward, negative downward, range −90°~+90°). The two points are time-ordered (left first), and once selected the whole line can be dragged or anchors fine-tuned;
Time range: drag A→B to frame two time points; the system draws a semi-transparent vertical band on the main chart (left/right borders + a top date-range label "start ~ end"). Anywhere inside the band is clickable — good for marking "event windows", "earnings/data release sessions", or "the time span of a completed structure".

Usage:

Angle for comparing slope changes: the angle difference between markup legs and pullback legs quantifies quickly — e.g., "this rally runs at 38° versus 52° for the last one, a clear slowdown" — more objective than eyeballing two lines. Note: the chart's horizontal axis (time) and vertical axis (price) have different dimensions; the angle is a screen-space value that changes with chart width/zoom — compare only within the same view;
Time range as "time anchors" for the chart: make the borders land on key time points (pattern start/end, major news moments), not a casual drag — a range's value lies in "what happened during this window".

5.8 Price Band

The price band is the horizontal twin of the time range: the time range paints "a period" as a vertical band; the price band paints "a price span" as a horizontal band across the whole chart.

Dragging: drag A→B (vertical span); the system normalizes the two points by price order (lower price on top edge, higher price on bottom edge) and draws a semi-transparent horizontal band — two border lines spanning the full width, each with a price label on the left (toFixed(2));
Typical use: marking high-volume zones / support-resistance bands / target-price zones — e.g., "64_000~64_300 below is the last two weeks' high-volume band; losing the lower edge means the bulls' line has failed". It expresses "a zone" better than a single horizontal line, and fits "care only about price, not the time window" better than a rectangle.

5.9 Fib Time Zones

Dragging: drag A→B (horizontal span); A is the origin and B only fixes the base length; boundaries fall at 1×, 2×, 3×, 5×, 8×, 13×, 21×, 34×, 55× the base from A (the first n=1 line solid, the rest dashed), with multiple labels on top;
Direction-sensitive: like cycle lines, the two points are order-preserving — A on the left extends right, A on the right extends left; flipping the drag direction mirrors the zones and won't auto-normalize away your intent;
Typical use: marking Fibonacci time windows — set A at the start of a significant trend and the base as "one complete swing"; the subsequent boundaries are the theoretical time nodes prone to regime changes/turns. Good for reviewing "did the last few highs/lows land near boundaries" and for pre-marking "where the next time window is" while a trend runs.

5.10 Layer Management: Hide, Lock, and Clear

Hide (👁/🚫): a hidden line disappears from the chart and no longer participates in any hit testing — it is merely "archived"; the stored data remains and it can be shown again anytime. Suits lines "not needed now but needed for later review" (e.g., the multiple parallel channels of some historical swing);
Lock (🔓/🔒): a locked line still renders (the structure stays on chart as reference) but can't be clicked or dragged — mouse/touch on it selects nothing, so no accidental drags. Suits key levels "already confirmed, reference only" (e.g., weekly-degree trend lines), protecting them from accidental bumps during daily zoom/pan;
Per-row delete (🗑): deletes just that one; clear all wipes every drawing of the current pair. Deletion is irreversible — think before clearing;
Row selection state: clicking a row in the panel selects the corresponding drawing on chart (blue + anchors), making it easy to locate a line from the list;
Entry: a "Layer Management" button (with current count) inside the "Drawings" collapsible panel on desktop/mobile; click to open the list.

5.11 Risk/Reward (R:R)

Drawing: click three points in sequence — A entry → B stop → C target (the order is the semantics; the system preserves order and never re-sorts); three horizontal lines spanning the full width appear: entry solid, stop/target dashed (all turn blue when selected), each with a price label on the left, and a payoff-ratio label (1:{ratio}) near the entry line on the right;
Algorithm: risk = |A − B| (entry to stop distance), reward = |A − C| (entry to target distance), ratio = reward / risk; when the stop equals the entry (risk=0), ratio records 0 without crashing — also a reminder that "a trade without a stop has no payoff ratio to speak of";
Typical use: before placing the order, lay out on the chart "where I exit if wrong, where I exit if right" and see at a glance whether R:R ≥ 2 (usually at least 1:2 to justify the risk); also good for review — losing trades are often not wrong on direction but on an inverted payoff ratio (taking trades where reward is far smaller than risk).

6. Shapes: Rectangle / Ellipse / Circle / Triangle / Arc

Shape tools are auxiliaries that "turn what the eye sees into a measurable object":

Tool	Use	Discipline
Rectangle	Frame a range/consolidation platform	Only the edges matter; don't trade inside
Triangle	Frame a converging/diverging pattern	The breakout direction matters more than the shape
Ellipse/circle/arc	Mark arc structures (rounding bottom/top)	Most subjective; review annotation only

Core reminder: the "shape" drawn by these tools only has operational value together with the confirmation signals of 04 · Advanced Candlesticks (breakout + volume + retest). A bare triangle outline says nothing.

7. Structure: XABCD Harmonics / Elliott Waves / Gann Angles

7.1 XABCD Harmonic Pattern

Usage: find patterns where XA, AB, BC, CD satisfy the classic ratios (e.g., AB ≈ 0.618 of XA, BC ≈ 0.382–0.886 of AB, CD ≈ 1.272–1.618 of BC), then look for reversal signals near D;
Controversy and traps:
- All five points are adjustable in hindsight; win rates are inflated in samples with "pattern first, price after";
- Always wait at D for "reversal confirmation" (reversal candle + volume); never counter-trend bottom-fish or top-pick just because "the ratios arrived";
Recommendation: treat it as a detector of "potential reversal zones", not a precise entry trigger.

7.2 Elliott Wave

Use the 5-point tool to mark the ends of waves 1–5 in order — write down your wave-count hypothesis before drawing;
Iron-rule check: wave 2 doesn't break wave 1's start, wave 3 isn't the shortest, wave 4 doesn't overlap wave 1 — any violation means recount;
The right mindset: wave counting is a "descriptive framework"; the tool's value is forcing you to think clearly about "where we are", not handing you deterministic buy/sell points.

7.3 Gann Fan

Nine angle lines from 1×8 to 8×1 out of an origin, representing the "time × price" speed scale:

The 1×1 line (45°) is the key reference: price above it → bullish strength; below → weakening;
Controversy: the angles depend on axis scaling (stretching the chart wider/taller changes all the angles), so Gann lines "morph" across software and zoom levels — always use them within the same coordinate frame, as a trend-speed reference only, never as precise signals.

7.4 Gann Box

The Gann box is the "zonal" version of the fan: drag A→B to frame a rectangle, and 10 angle lines are auto-drawn inside — 1×1 double diagonals (bold) + 1×2 / 2×1 from the four corners.

Usage: after framing a consolidation range, watch price's "angle rhythm" inside the box — riding 1×1 = equilibrium; losing it and switching to 2×1 support = deceleration; steep climb along 2×1 = acceleration leg (beware the emotional climax);
Resonance: spots where the box's 1×1 diagonal / upper edge overlaps Fibonacci 0.5/0.618 or channel boundaries carry more reference value;
For the full practice (including time-price balance and common mistakes), see 08 · Gann Box and Angles in Practice.

8. Annotation: Measure and Text

8.1 Measure

The measure tool shows the price gap and percentage change between two points; best for:

Measuring "breakout height": from the box top to the target, for estimating the equal-move objective after a break;
Measuring "pullback depth": quickly judging whether the current pullback is 0.382 or 0.5, cross-checked with Fibonacci;
Measure gives numbers, not opinions — numbers only mean something inside your trading plan.

8.2 Text

9. The Discipline of Drawing: From "Can Draw" to "Can Use"

9.1 Three Iron Rules for Anchor Selection

Anchors must come from significant structure: prior highs/lows, volume-heavy tops/bottoms, long-consolidation boundaries — never random swing extremes;
Fewer anchors, more credibility: a trend line needs 3+ tests to have value; frequent redraws mean you are "accommodating the market";
Multi-timeframe alignment: a line drawn on the daily should still "mean something" on 4H/1H — big-timeframe anchors stay valid on smaller ones, not vice versa (method in 04 · Advanced Candlesticks).

9.2 Every Line Must Have an "Invalidation Condition"

After drawing any line, ask yourself three questions:

Question	Consequence of having no answer
What view does this line represent?	It's decoration — delete it
What do I do when price hits the line?	No trigger plan = not drawn at all
What price action proves this line wrong?	No invalidation condition = you will hold losers forever

💡 The Ultimate Test of Drawing

9.3 Common Mistakes Checklist

❌ Over-drawing (more than ~10 lines per screen): information overload equals no information;
❌ Repeatedly moving anchors to "fit" the market: this is drawing-tool overfitting — the same disease as the backtest overfitting covered in 06 · Critique and Validation;
❌ Treating retracement/target levels as exact prices: they are zones, not points;
❌ Replacing risk control with drawing signals: however good the structural call, the stops/position rules of the 07 · Trading Systems chapter are the backstop.

9.4 Mobile & Touch Interaction Differences

Touch has no hover state, so the interaction mindset differs from desktop — four differences worth knowing:

Two-step editing: desktop separates select vs. edit via "hover + click"; touch switches to "tap to select → drag to edit" to avoid accidental touches;
Auto-return to "mouse" after each drawing: on mobile, every completed drawing (including text confirmation) switches the tool back to "mouse" — right after drawing, the finger most likely wants to fine-tune, not draw another; leaving the tool active invites accidental creation; desktop keeps continuous drawing;
Double-click / long-press to rewrite text: desktop double-clicks the text body, mobile long-presses it (~250ms) to open the editor directly — no panel needed; works only in "mouse" read-only mode, so it never clashes with drawing gestures;
Reading data in comfort: after dragging or long-press pinning, the crosshair lingers about 2 seconds on release (a tap or any new gesture clears it at once), and the OHLC tooltip flips at container edges to stay fully inside the viewport — even drags in the mid-lower chart can read the last row.

10. Summary: The "Dao" and "Shu" of Drawing

Level	Content	Corresponding tools
Shu (how to draw)	Anchors from significant structure, multi-point confirmation, multi-timeframe alignment	Trend lines/channels/Fibonacci
Fa (how to use)	Hypothesis first, zones not points, explicit invalidation conditions	Horizontal lines/XABCD/measure
Dao (why)	Drawing is the visualization of thinking, not a prediction grail	All tools

Read this chapter's discipline together with 07 · Trading Systems and 08 · Pitfalls; only then does drawing graduate from "decoration" to "decision".

Conventions

All prices and figures in this article are illustrative and represent no specific instrument or market.
The exact interactions of each drawing tool (anchor count, drag method) are governed by the actual Kline Buty interface.
Fibonacci ratios and harmonic-pattern ratios are "statistical tendencies", not laws; before use, understand the validation methodology of 06 · Critique and Validation.
For counter-trend trades (e.g., harmonic D-point bottom-fishing), be sure to read the margin and liquidation sections of the 03 · Futures chapter first.

⚠️ Risk Warning