LEARN ELLIOTT WAVES
Rules for checking a wave count.
A rule is a condition required by a particular pattern. A guideline describes a tendency. Check the pattern first, then its price relationships and internal structure.
Define the pattern being checked.
The checks below concern an ordinary upward impulse at one degree. Reverse the price directions for a downward impulse. They are a first screen, not the full definition of every Elliott pattern.
Diagonals have their own structure and can include wave 1/4 overlap. Calling a failed impulse a diagonal requires evidence for that separate pattern. The source also discusses rare futures-price extremes; this lesson uses the ordinary non-overlapping case.
Three introductory price checks.
| Check | Invented example |
|---|---|
| Wave 2 must not completely retrace wave 1. | Start 100, wave 1 ends at 110: wave 2 at 104 passes this check; 98 fails it. |
| Wave 3 must not be the shortest of waves 1, 3 and 5. | Lengths 10, 20, 14 pass this check. Lengths 10, 6, 9 fail because the third is shortest. |
| In an ordinary impulse, wave 4 must not enter wave 1 price territory. | Wave 1 ends at 110: wave 4 at 116 passes this check; 108 overlaps and fails. |
Endpoints are only the beginning.
Wave 3 must advance beyond wave 1. An impulse has 5–3–5–3–5 subdivisions; wave 3 is itself an impulse. The child structures must fit the parent’s boundaries and roles. If the data do not reveal them, keep that part unconfirmed.
Do not assume a fifth wave is finished simply because the chart ends. Until the relevant components are sufficiently developed, a full comparison of completed waves is not available.
Do not turn a tendency into a rule.
Alternation suggests differences between waves 2 and 4. Fibonacci relationships can help compare proportions. Neither a preferred ratio nor a visually attractive count substitutes for the required structure.
Common mistakes are treating wave 3 as always the longest, mixing degrees, ignoring an inconvenient extreme and moving the start only to rescue a count. In the exercise, a 60% wave 2 retracement is not invalid merely because it differs from 61.8%. Record the model, the failed condition and the next evidence needed.
Theory references
Check the rules with BTC data
See the three price checks in the BTCUSDT example. Then inspect wave subdivisions: passing these checks does not by itself confirm an impulse.
Elliott wave time cycles: measurements, guidelines and rules
A time measurement answers how long a selected movement lasted. A timing hypothesis asks whether another movement might have a related duration. Neither, by itself, identifies a valid wave pattern or guarantees a reversal date. Begin with the proposed structure, wave degree and data cutoff; then record the timing comparison separately.
Keep the source of a timing condition visible
Elliott Wave International’s Fibonacci overview includes timing among the uses of ratios. Its rules and guidelines explanation distinguishes mandatory impulse conditions from interpretive guidelines. A selected time ratio cannot repair a violation of a price or subdivision rule.
Glenn Neely’s description of NEoWave presents additional rules and post-pattern tests, including time conditions, as extensions of orthodox Elliott wave analysis. Keep a condition from that method attached to its author, pattern and version. Do not transfer a minimum duration from one method or pattern into every Elliott count. This lesson does not adopt a universal minimum time for wave 2 or wave 4.
A synthetic timing worksheet
The following numbers are invented bar indices, with no market prices. A, B and C are measurement anchors, not labels for an A–B–C correction. Suppose A is at index 100, B at 121 and the comparison starts from C at 134. Count elapsed bar intervals: d = 121 − 100 = 21. There are 22 indexed bar positions when both endpoints are included; that is a different quantity.
| Chosen comparison | Arithmetic | Nearest whole interval | Position measured from C |
|---|---|---|---|
| 0.618 × reference duration | 21 × 0.618 = 12.978 | 13 | 134 + 13 = 147 |
| 1.618 × reference duration | 21 × 1.618 = 33.978 | 34 | 134 + 34 = 168 |
The ratios and rounding convention were chosen for this exercise. The positions 147 and 168 are places to inspect under that hypothesis, not promised turning points. They do not establish that a named wave must last that long, that a recurring cycle exists, or that a pattern is complete. If a real study uses a tolerance window, choose and record its width before viewing the outcome.
Bars, sessions and calendar time are different clocks
Write down the instrument, feed, timeframe, session boundary and time zone. Daily trading bars skip closed sessions; calendar days do not. Missing bars and different session definitions change a bar count. If two turning points fall inside the same bar, that bar alone may not reveal their order or elapsed time. Do not convert a projected index into a date until the relevant trading calendar and counting convention are defined.
For each comparison, save the exact start and end anchors, the reference duration, the ratio, rounding and tolerance, the information available at the cutoff, and what would reject the hypothesis. Keep completed, developing and unresolved structures distinct. Use alternative counts to preserve other valid interpretations; the Fibonacci lesson explains the separate price calculations. A timing observation adds context to a structure check; it supplies no measured trading probability.
Study offline with the original PDF guide
Rules, counting, corrections and Fibonacci, with invented worked examples and an answer key. Available in this page language.