05 · Elliott Waves · Gann · Chan Theory
Elliott Wave, Gann theory, and Chan Theory are the three tallest peaks of "mysticism density" in the technical analysis world: logically self-consistent, complete systems with huge followings, yet almost never positively verified by academic research and almost never adopted by professional institutions. This article hypes nothing and trashes nothing — it lays out the principles, rules, and controversies of all three, then delivers their real merit and correct usage.
💡 Master Principle
Remember one master principle first: these theories work as "a language for describing market structure", not as "tools for predicting the market". Used to explain the past, they are almost always right; used to predict the future, it's a coin flip.
1. Elliott Wave Theory
1.1 Core Structure: 8 Waves = 5 Impulse + 3 Corrective
Ralph Elliott proposed in the 1930s that market movement alternates between impulse waves and corrective waves; a complete cycle is 8 waves:
Complete cycle (8 waves):
5 (new high)
╱╲
4╱ ╲
3 ╱╲ ╲
╱ ╲ ╲
╱ 2 ╲ ╲
╱╲ ╲ ╲ B
1 ╲ ╲A ╲╲ ╱
╲ ╲ ╲ ╱
╲ ╲C ╲╱
╲ ╲
╲
╲
1–5 = impulse waves (with the main trend, 5 waves)
A–C = corrective waves (against the main trend, 3 waves)- Impulse waves (1, 2, 3, 4, 5): in the main trend's direction; wave 3 is usually the longest and fiercest;
- Corrective waves (A, B, C): against the main trend; wave B is often read as "the bulls' last struggle";
- Nesting into larger degrees: after the 5+3=8 waves complete, the market moves into the next leg of a larger-degree 5-wave impulse — waves are fractally nested; every 5-wave impulse contains smaller 5+3 structures, nested ever deeper (one root of the "a thousand people, a thousand wave counts" problem below).
1.2 Iron Rules and Guidelines
Elliott Wave has 3 "iron rules" (violation invalidates the count) and several "guidelines" (tendency judgments):
| Class | Content | Note |
|---|---|---|
| Iron rule 1 | Wave 2 cannot retrace past the start of wave 1 | Otherwise the entire count is void |
| Iron rule 2 | Wave 3 cannot be the shortest of the three impulse waves | Wave 3 is usually the longest with the heaviest volume (the "markup wave") |
| Iron rule 3 | Wave 4 cannot overlap wave 1's price range | Exceptions allowed only in rare ultra-strong trends |
| Guideline | Wave 4 tends to be a sideways correction, wave 2 a steep one | The two should "alternate" in shape |
| Guideline | If wave 1 and wave 4 overlap, the count can be revised | The flexibility = the subjectivity |
Fibonacci relationships: Elliott Wave is often bound to Fibonacci ratios — wave 2 often retraces 0.5/0.618 of wave 1, wave 3 ≈ 1.618× wave 1, wave 4 often retraces 0.382 of wave 3, wave 5 ≈ wave 1 in length. Note: these are statistical tendencies that "often occur", not laws — the market does not run to satisfy Fibonacci.
1.3 The Subjectivity Controversy of Wave Counting
- Counting in hindsight is always right: once the move is over, anyone can draw a perfect 5-3-8 structure — but on the same chart, different analysts count completely different waves (where exactly wave 1 starts, whether wave 4 is a triangle or a flat, whether wave B has ended) — that is "a thousand people, a thousand wave counts";
- The plasticity of waves: the rules leave too much room for "exceptions" and "revisions" (alternation, extensions, truncations); almost any price path can be bent into some legitimate count — a theory that is always "right" is one that can never be falsified;
- Iron rules only filter extreme errors, they don't guarantee correctness: several mutually contradictory counts can all satisfy the 3 iron rules (the same price can be the end of wave 3 or the start of wave 5).
2. Gann Theory
2.1 Who Was Gann
William D. Gann (1878–1955), an American speculator. He did make money (though some accounts say his late-life trading record was unimpressive) and left behind a large body of theory centered on "time + price" with heavy mystical overtones. Note: Gann himself left almost no reproducible trading documentation; the surviving "Gann theory" is mostly second-hand reconstruction by later compilers.
2.2 The Three Core Tools
| Tool | Content | Principle |
|---|---|---|
| Gann angles | 1×1, 1×2, 2×1 etc. rays drawn from key lows/highs, representing fixed price-to-time ratios | Price × time equilibrium: 1×1 is 45° (one unit of price per unit of time) |
| Time cycles | Markets tend to turn at specific intervals (7 days, 30 days, 90 days, anniversaries) | Based on astronomical cycles and natural number sequences |
| Square theory | Dividing price by the 360° circle/square (0.25/0.5/0.75 etc.) to find "geometric correspondences" of support/resistance | The philosophy that all is number and cycles repeat |
Gann angles sketch (from a low):
price│ ╱1×1 (45°, equilibrium)
│ ╱
│ ╱1×2 (price outruns time, strong)
│ ╱
│ ╱ 2×1 (time dominates, weak)
│ ╱
│╱
└──────────────────→ time2.3 Philosophical Background and the Verification Impasse
- Philosophical background: Gann was deeply influenced by the "law of vibration" and by theology/astrology, holding that price vibration shares its source with natural laws (seasons, planetary cycles) and therefore markets can be derived geometrically and numerically. This system has the conclusion first (cycles repeat), the method after (angles/sequences);
- Why it is nearly impossible to verify:
- Subjective tool selection — the anchor of an angle line (which low) and the ratio (1×1 or 1×4) are chosen by the drawer; different people draw completely different lines;
- Vague "time window" definitions — saying "July is a turning month" cannot be falsified: July has 31 days, and a turn on any one of them counts as "confirmed"; the theory keeps too many escape hatches;
- Parameters disconnected from the real rules — the algorithms in Gann's original notes were never fully published; textbooks contradict each other, so no unified standard for backtesting can even be built;
- Survivorship narrative — Gann's legend rests on "prediction cases" in his books, while failed predictions were never fully recorded: a classic selection bias.
3. Chan Theory
3.1 What "Chan Zhong Shuo Chan" Is
"Chan Zhong Shuo Chan" ("Entanglement" / Chan Theory) is a stock-market theory serialized on a blog in 2006–2008 (the author's pen name was "Chan Zhong Shuo Chan"; he died in 2008). It has a large following in the A-share community and is the most influential natively Chinese technical analysis system. Four core building blocks:
| Block | Definition | Role |
|---|---|---|
| Stroke (bi) | The connection between adjacent top and bottom fractals (top fractal + bottom fractal = one stroke) | The minimal unit of analysis |
| Segment (xianduan) | At least three strokes, with directional continuation | The upgraded structure above strokes |
| Pivot (zhongshu) | The overlapping price zone where price oscillates repeatedly (at least three sub-degree moves overlapping) | The "base camp" of bull-bear balance |
| Divergence (beichi) | Price makes a new high/low but momentum (MACD area/bar height) does not | The criterion for reversal |
Chan Theory structure sketch:
price│ ╱╲ ╱╲
│ ╱ ╲ ╱ ╲
│ ╱ ╲╱ ╲
│ ╱ ┌─pivot─┐ ╲ ← zone where price repeatedly overlaps
│ ╱ └─────┘ ╲
│╱ (stroke) (turn after divergence)
└──────────────────→ time3.2 The Degree Concept and the Three Buy Points
- Degrees: Chan Theory stresses "every move must complete" — a move at any degree consists of "consolidation + trend", and larger degrees are recursively built from smaller ones. You must fix the degree first (which degree's pivots and divergences you trade by); degree confusion is the most common beginner mistake;
- The three buy points (sell points are exact mirrors):
- First buy: divergence appears at the last pivot of a downtrend → the buy point of the trend's end (bottom-fishing in nature; hardest, lowest win rate);
- Second buy: after the first buy, the pullback holds above a new low → the second-low buy point (steadier);
- Third buy: after breaking out of a pivot, the retest fails to re-enter the pivot → the trend-continuation buy point (higher win rate, suits right-side trading).
3.3 Popularity in the A-Share Circle and the Subjectivity Analysis
- Why it spread: born in the great 2005–2007 A-share bull market, packaged and promoted by countless "battle methods"; its "geometric, structural" language is more precise in form than traditional candlestick theory, it was freely published, and it satisfied retail traders' psychological need for "a complete system";
- Its subjectivity is just as severe: stroke division has threshold disputes (when does a fractal count), segment rules are complicated, and pivot-degree determination relies on recursion — on the same chart, different Chan practitioners still draw different strokes, segments, and pivots, hence the saying "a thousand people, a thousand Chans";
- Unverifiable: Chan Theory lacks unified backtest definitions (stroke/segment/pivot determination needs many parameters); serious academic or quantitative research on its effectiveness is essentially absent; the "divergence" criterion fails repeatedly in ranging markets.
4. The Shared Critique of All Three
4.1 Great at Explaining the Past, Weak at Predicting the Future
| Critique | Manifestation |
|---|---|
| Perfect in hindsight | Once the move is over, the waves can be counted, the angle lines drawn, the pivots divided — everything looks "right" |
| Vague in real time | While the move runs, "which wave/segment are we in" is never settled; judgments contradict each other |
| Unfalsifiable | Each theory builds in enough "exception clauses" (extensions, alternation, time windows) to explain any outcome |
| Results depend on the drawer | Anchors, parameters, and degrees are all subjective choices; the same data supports several "correct" conclusions |
4.2 A Thousand People, a Thousand Waves / Chans
The same move, three analysts, three structures:
Analyst A: wave 3 is running, target Fibonacci 1.618× → keep holding
Analyst B: wave 5 is done, a large correction is due → go flat, watch
Analyst C: the count failed, this is the start of an A-B-C correction → flip short
→ All three have "complete, self-consistent" reasons — but at most one is right- Division rules too complex to reproduce is the shared fatal flaw: the more complex the rules, the larger the explanation space and the worse the predictive precision;
- This is textbook overfitting: with 100 rules to "fit" history, some set always fits — but that set has zero binding force on the future.
4.3 The Gap from Scientific Methodology
The value of a scientific theory lies in being falsifiable and reproducible: state clear prediction conditions and let the market test them. Elliott/Gann/Chan never completed this step — they are closer to "religious explanation frameworks": believers use them to explain everything, skeptics see ambiguity. "Explains everything" and "predicts nothing" are two sides of the same coin.
5. A Realistic Assessment of Their Merit
5.1 Why Professional Institutions Almost Never Use Them
- Unquantifiable: institutional strategies need backtestable, parameterizable rules; strokes/pivots/wave counts carry too much subjectivity to enter a quant pipeline;
- No academic support: academic research on the effectiveness of Elliott/Gann/Chan is nearly blank — not because nobody studies them, but because not even a "unified determination standard" can be established;
- Unfalsifiable means uniterable: an institution's edge comes from "using data to eliminate dead methods"; a theory that is always "right" can never be improved;
- Substitutes exist: the same "structure description" needs can be met with rule-based tools — trend lines, moving-average systems, channels, MSS (market structure shift) — which at least can be backtested.
5.2 Their Value as "Thinking Frameworks"
From another angle: while none of the three works as a prediction tool, each contributes a useful lens —
| Theory | Contributed framework | What you can absorb |
|---|---|---|
| Elliott Wave | "Trends have structure; movement has rhythm" | The impulse/correction segmentation mindset; awareness of momentum exhaustion late in trends (echoes top divergence) |
| Gann theory | "Time and price confirm each other" | Watch the crossings of time windows and key price levels, used with indicators rather than alone |
| Chan Theory | "Moves are pivot-driven; divergence finds turns" | Use pivots to describe ranges; use divergence (momentum exhaustion) to assist reversal judgment |
💡 Positioning Summary
Positioning summary: they are "description languages of market structure", not "holy grails". Use them to raise your observational resolution (segment moves more finely, treat divergence and exhaustion more alertly) — not to tell fortunes.
6. How to Use Them
6.1 As a Descriptive Language, Not a Prediction Tool
- Use the wave "impulse/correction" to describe: are we in a markup wave or a correction (the trend-vs-range judgment);
- Use Chan pivots to replace the "box" concept: price inside a pivot = don't act; leaving the pivot + a retest that holds = a trend opportunity;
- Use Gann's "time + price" crossings to remind yourself: when a key price level appears near a key time window, raise alertness and tighten stops — not to predict the turn by it.
6.2 Combined with Volume-Price and Indicators
| Scenario | Correct usage | Wrong usage |
|---|---|---|
| A "wave 5" is counted | Wait for independent confirmation (top divergence + expansion with stalling) before reducing | Shorting the moment wave 5 is counted — wave 5 can extend again and again |
| A Chan "first buy" appears | Probe with a small position + strict stop-loss + add on the second buy | Going heavy at the first buy — divergence can diverge again |
| A Gann time window nears | Reduce position, tighten stops, mostly stand aside | Heavily pre-betting the turn inside the window |
6.3 Always Let the Stop-Loss Catch You
- Any prediction from any theory is only a "possibility": the wave count can be wrong, the time window can fail, divergence can freeze — that's the norm, not the exception;
- The moment price negates a structural call, admit it immediately: e.g., "wave 2 broke the start of wave 1" — however pretty the count was, the structure is broken; stop out;
- Position sizing and stops follow the rules of the 07 · Trading Systems chapter: the theory only decides "where to watch"; position management decides "survive to the next opportunity";
- Beware "theory faith": the more complex the theory, the more easily it self-persuades — and self-persuasion is the prelude to heavy losses — when a theory makes you feel "very confident", that is exactly when to check whether you've fallen into confirmation bias.
⚠️ Risk Warning
Structural theories like Elliott, Gann, and Chan guarantee no returns, and no serious academic evidence shows their predictive power beats randomness — their biggest risk is precisely that their completeness, self-consistency, and perfect hindsight easily breed the illusion of "I have mastered the market's law", leading traders to abandon position management and stop discipline: the classic path of retail blow-ups. Treat these three as auxiliary frameworks for raising observational resolution; any prediction based on them requires independent volume-price/indicator confirmation, with the stop-loss as the final backstop. This article is for educational purposes only, does not constitute investment advice, and certainly does not encourage leveraged casino-style trading in any form.